Clinical experiences with optical coherence tomography in epithelial (pre)malignancies

(1)

Ronni Wessels

Clinical experiences with optical coherence

tomography in epithelial (pre)malignancies

(2)

(3)

linical eǆperiences ǁith opƟcal coherence tomography in

epithelial (pre)malignancies

(4)

ISBN: 978-94-6259-681-8

>aǇ-oƵt Θ over: Ilse StronŬs͕ ƉersoonlijŬƉroeĨsĐhriŌ͘nl Printed by: Ipskamp Drukkers BV, Enschede, The Netherlands

The content oĨ this thesis has been approved by ProĨ͘ dr͘ T͘:͘D͘ Ruers and ProĨ͘ dr͘ ͘'͘:͘D͘ van >eeuwen͘

The publicaƟon oĨ this thesis was Įnancially supported by Twente hniversity - TNW Ĩaculty, hipSoŌ and Santec orporaƟon͘

ll riŐhts reserved͘ No part oĨ this thesis may be reproduced or transmiƩed in any Ĩorm or by any means, electronic or mechanical, includinŐ photocopy, recordinŐ or any inĨormaƟon storaŐe or retrieval system, without prior wriƩen permission oĨ the copyriŐht owner͘

(5)

CLINICAL EXPERIENCES WITH OPTICAL

COHERENCE TOMOGRAPHY IN EPITHELIAL

ΈPREΉMALIGNANCIES

PROEFSCHRIFT

ter verkrijŐinŐ van

de Őraad van doctor aan de hniversiteit Twente, op ŐeǌaŐ van de Rector DaŐniĮcus,

ProĨ͘ dr͘ H͘ Brinksma,

volŐens besluit van het ColleŐe voor PromoƟes in het openbaar te verdediŐen op donderdaŐ 21 mei 2015 om 12͘45 uur

door

Ronni Wessels Őeboren op 26 Ĩebruari 1984

(6)

PROMOTIECOMMISSIE

Promotores: ProĨ͘ dr͘ T͘:͘D͘ Ruers

ProĨ͘ dr͘ ͘'͘:͘D͘ van >eeuwen Co-promotores: Dr͘ ir͘ D͘:͘ Faber

Dr͘ D͘ Van Beurden

OveriŐe leden: ProĨ͘ dr͘ ir͘ :͘W͘D͘ HilŐenkamp ProĨ͘ dr͘ ir͘ C͘H͘ Slump

ProĨ͘ dr͘ ir͘ H͘:͘C͘D͘ SterenborŐ Dr͘ S͘ Danohar

Dr͘ '͘N͘ Relyveld Dr͘ C͘͘R͘ >ok

(7)

'eneral IntroducƟon

Thesis DoƟvaƟon and im

Thesis Structure

(10)

Chapter 1 8

GENERAL INTRODUCTION

Dore than 80й oĨ all cancers oriŐinate Ĩrom the epithelium ΀1΁͘ Epithelium lines orŐans and caviƟes inside the body and Ĩorms the skin on the outside͘ Epithelium reŐenerates constantly and the Ɵssue is prone to the development oĨ cancer ΀2΁͘ In this thesis we concentrate on cancers arisinŐ in the epithelial lininŐ oĨ the male and Ĩemale Őenital area and oĨ the skin͘ In the last 30 years, the incidence of vulvar intraepithelial neoplasia (VIN) has increased more than 400й to approǆimately 2͘5 cases per 100,000 women in the hnited States ΀3΁͘ In the Netherlands, the incidence of VIN was 2͘2 per 100,000 women a year in 2005 ΀4΁͘ VIN is a premaliŐnant skin disorder that oŌen causes pruritus, pain, and psychoseǆual dysfuncƟon͘ It is diaŐnosed throuŐh painful punch biopsy ΀5΁͘ VIN was previously Őraded into VIN 1 to 3͘ Recently, a new classiĮcaƟon was adapted, which divides VIN into diīerenƟated-type and usual-type VIN (dVIN and uVIN) ΀6΁͘ dVIN is associated with lichen sclerosis and occurs more oŌen in elderly women͘ In younŐer women, a persistent infecƟon of human papillomavirus (HPV) ΀7-8΁, parƟcularly HPV type 16 and 18, is associated with the occurrence of uVIN͘ The only way to obtain deĮnite diaŐnosis of VIN in case of a vulvar lesion of uncertain siŐniĮcance is by takinŐ a punch biopsy, which can be painful ΀5΁͘ In some paƟents mulƟple biopsies are needed to diaŐnose VIN, as VIN can recur oŌen͘ non-invasive diaŐnosƟc opƟcal tool that can diīerenƟate between normal vulvar Ɵssue and VIN, would therefore be convenient͘ Treatment of VIN consists of surŐical eǆcision, laser vaporiǌaƟon or medical therapy͘ However, every aƩempt is made to avoid vulvar muƟlaƟon that may possibly lead to psychoseǆual distress ΀9΁͘ Nevertheless, even aŌer treatment there is a chance of recurrence of VIN or occult invasion͘ Therefore, paƟents are reŐularly eǆamined to foresee occult invasion and check for possible new VIN lesions ΀10΁͘ Both types of VIN may proŐress into invasive vulvar sƋuamous cell carcinoma (VSCC)͘ The proŐression rate of VIN into VSCC is about 9й in untreated paƟents, and 3͘3й in paƟents aŌer treatment ΀10΁͘

Treatment of VSCC consists of a wide local eǆcision of the tumour with lymphadenectomy of the inŐuino-femoral areas with separate incisions, toŐether with radiotherapy andͬor chemotherapy for locally advanced or recurrent disease ΀11΁͘ Standard surŐical marŐins for local treatment of paƟents with VSCC should be at least 1 cm ΀12-14΁͘ For eǆample, it is known that histoloŐical marŐins around the VSCC of less than 8 mm will result in local recurrences up to 50й͘ The challenŐe in treatment of VSCC is to keep surŐical marŐins larŐe enouŐh to minimiǌe the chance of recurrence, but at the same Ɵme, preserve as much Őenital Ɵssue as possible to diminish seǆual and psycholoŐical morbidity ΀15΁͘ These important resecƟon marŐins are determined by the ŐynaecoloŐist with ͚naked eye͛ assessment durinŐ surŐery͘ n opƟcal tool that is able to disƟnŐuish diīerent Ɵssue types from one another, could be helpful in determininŐ resecƟon marŐins durinŐ surŐery͘

In men, similar challenŐes are present in diaŐnosinŐ and treaƟnŐ penile intraepithelial neoplasia and penile sƋuamous cell carcinoma͘ In the Western world penile cancer is rare, but in some

(11)

'eneral IntroducƟon, DoƟvaƟon and im, Thesis Structure 9

1

frican, sian and South merican countries its incidence rate can be up to 10й͘ In ParaŐuay and hŐanda incidence rates are 4͘2 and 4͘4 respecƟvely per 100,000 men ΀16-17΁͘ These numbers are in contrast to Western Europe and the hnited States where aŐe-standardiǌed incidence rates ranŐe from 0͘3 to 1͘0 per 100,000 men and penile cancer accounts for only 0͘4-0͘6й of all maliŐnancies ΀16, 18΁͘ Dore than 95й of penile tumours are sƋuamous cell carcinomas (SCC) ΀19΁͘ SCC can be preceded by penile intraepithelial lesions (PIN) ΀20΁͘ The term PIN covers diīerent clinical presentaƟons of dysplasƟc lesions͘ The morpholoŐy of PIN lesions can vary Őreatly: from piŐmented to leukoplasƟc papules, erythroplasic macules and keraƟniǌed condylomas͘ HistopatholoŐically, PIN consist of disorŐaniǌed basal and parabasal layers combined with cellular atypia, eǆhibiƟnŐ atypical mitosis ΀21΁͘ Before treatment starts, PIN is diaŐnosed throuŐh painful punch biopsy͘ To diminish the burden of biopsies, an opƟcal non-invasive imaŐinŐ tool that can diīerenƟate between healthy penile skin and PIN, would be beneĮcial͘ The Őoal of treatment of PIN is, similarly to treatment of VIN in women, to eradicate the disease while limiƟnŐ penile muƟlaƟon͘ Diīerent treatments are available, from topical therapy like podophyllotoǆin, imiƋuimod, to surŐical methods as laser ablaƟon and eǆcision͘ No randomiǌed clinical trials are known comparinŐ various treatment modaliƟes͘ The choice of treatment is Őenerally based on preferences and skills of the uroloŐist ΀22΁͘ non-invasive imaŐinŐ techniƋue could also be used in the follow-up of a treated lesion͘

Skin cancer can occur anywhere on the skin, not only in the Őenital area͘ There are two diīerent kinds of skin cancer: melanoma, and non-melanoma skin cancer͘ Non-melanoma skin cancer (NSDC), includinŐ basal cell carcinomas (BCC), sƋuamous cell carcinomas (SCC) and the premaliŐnant acƟnic keratosis (<), is the most prevalent cancer in liŐht-skinned populaƟons ΀23΁͘ Some paƟents suīer from mulƟple lesions durinŐ life, others suīer from lesions at locaƟons where resecƟon results in bad cosmeƟc outcome ΀24΁͘ These cancers have a beniŐn character͖ they do not inŇuence survival, in contrary to melanoma͘ Cutaneous melanoma is much more aŐŐressive͘ Per year 55,489 paƟents die worldwide of melanomas of the skin ΀25΁͘ To reduce mortality resulƟnŐ from melanomas, recoŐniƟon of this maliŐnancy by clinicians in an early staŐe is important ΀26΁͘ DermatoloŐist only have a few tools to assist with the diaŐnosis of melanoma͘ Primarily they rely on their clinical judŐment to decide whether to perform eǆcision of a piŐmented lesion͘ The clinicians͛ Őoal is to detect melanomas with the hiŐhest accuracy, while avoidinŐ unnecessary eǆcisions͘ hnaided (naked eye) diaŐnosis of maliŐnant melanoma by a trained dermatoloŐists has shown a diaŐnosƟc accuracy of around 60й ΀27΁͘ non-invasive diaŐnosƟc tool that is able to diīerenƟate between beniŐn piŐmented lesions and maliŐnant melanomas would be helpful͘ PaƟents would beneĮt from a non-invasive imaŐinŐ tool that is able to obtain a non-invasive ͚opƟcal biopsy͛ and determine marŐins of eǆcision, especially in areas where resecƟon is muƟlaƟnŐ͘

The most accepted noninvasive techniƋue that is beinŐ used clinically for melanoma detecƟon is dermoscopy (also known as dermatoscopy or epiluminescence microscopy)͘ dermoscope

(12)

Chapter 1 10

consists of a maŐniĮer (usually ǆ 10), a non-polariǌed liŐht source and a liƋuid medium between the skin and the dermoscope͛s Őlass slide͘ The dermoscope eliminates liŐht reŇecƟon ΀28΁͘ In this way, it provides clinicians the ability to look deeper into the piŐment and vascular structures in the skin in vivo͘ Several alŐorithmic methods have been developed to improve the diaŐnosis of melanocyƟc lesions with the use of dermoscopy ΀29΁͘ The three most used alŐorithmic methods are ƋualitaƟve paƩern analysis, the BCD-rule (where сasymmetry, Bсborder, Cсcolor and Dсdiīerent dermoscopic structures) and the 7-point checklist where four major criteria (for eǆample: atypical piŐment network) and three minor criteria (for eǆample: reŐression paƩern) are scored ΀29΁͘ When used by a trained clinician, dermoscopy increases the diaŐnosƟc accuracy for melanoma ΀27΁͘ Confocal laser scanninŐ microscopy (C>SD) (also known as reŇectance confocal microscopy (RCD)) is another techniƋue that is increasinŐly beinŐ studied ΀30΁͘ This techniƋue enables en face imaŐes with ƋuasihistoloŐical resoluƟon͘ The confocal microscope uses a near-infrared laser with a wavelenŐth of 830 nm͘ >ateral resoluƟon and aǆial resoluƟon of 1-2 ђm and 3-5 ђm respecƟvely are reached͘ ImaŐinŐ depth is 200-500 ђm, which allows for visualiǌaƟon of the epidermis and superĮcial dermis at cellular resoluƟon ΀31΁͘ Based on several features that can be disƟnŐuished in C>SD imaŐes, diīerent alŐorithms were developed ΀32-34΁͘ C>SD appears to be a promisinŐ techniƋue for melanoma diaŐnosis ΀33, 35-37΁͘ ThouŐh, an important limitaƟon of C>SD in its current state, is the limited imaŐinŐ depth (approǆimately 250-500 ђm)͘ Therefore, processes beneath the papillary dermis cannot be evaluated reliably͘ Dermoscopy and C>SD enable addiƟonal informaƟon about melanocyƟc lesions, thouŐh both techniƋues lack suĸcient depth imaŐinŐ͘ In addiƟon, dermoscopy and CS>D Őive ƋualitaƟve informaƟon about lesions͘ It is not possible to obtain a ƋuanƟtaƟve parameter that can help diīerenƟaƟnŐ between nevi and melanoma͘ In conclusion, paƟents with these epithelial cancers would beneĮt from an imaŐinŐ techniƋue that is able to diīerenƟate between dissimilar Ɵssue types͘ This imaŐinŐ techniƋue would provide a non-invasive ͚opƟcal biopsy͛ to a certain depth that could be helpful in diaŐnosinŐ PIN, VIN and melanoma͘ In addiƟon, a non-invasive imaŐinŐ techniƋue could help in determininŐ safe marŐins of eǆcision durinŐ surŐery of penile and vulvar sƋuamous cell carcinoma and durinŐ eǆcision of melanoma lesions͘

Imaging techniƋƵes

variety of imaŐinŐ modaliƟes able of imaŐinŐ the human body are available͘ ll these techniƋues eǆperience one major limitaƟon: with increasinŐ depth, resoluƟon diminishes (ĮŐure 1)͘

SinŐle photon emission computed tomoŐraphy (SPECT) and positron emission tomoŐraphy (PET) use radioacƟve tracer isotopes that are injected into the body͘ These tracers are incorporated in cells in the body and emit a positron as the radioisotope underŐoes positron emission decay͘ PETͬSPECT allow imaŐinŐ deep in the body and whole body imaŐinŐ is possible in this way͘ On

(13)

1

the other hand PETͬSPECT imaŐes lack spaƟal resoluƟon and anatomic correlaƟon͖ therefore these imaŐes are combined with imaŐes made with CT (PET-CT)͘ Computed tomoŐraphy (CT) is used to imaŐe the human body with aǆial ͚slices͛ that shows the anatomic morpholoŐy really well͘ DaŐneƟc resonance imaŐinŐ (DRI) uses electromaŐneƟc Įelds and is hiŐhly sensiƟve for hydroŐen in the cells͘ hltrasound imaŐinŐ uses pulsed sound waves to Őather imaŐes of the human body͘ No harmful radiaƟon is used, thouŐh direct contact with the skin is needed to create the imaŐes͘

CT, DRI and ultrasound are able to imaŐe orŐans very well, thouŐh the resoluƟon of these imaŐinŐ techniƋue lacks the ability to look into Ɵssue at cellular resoluƟon͘ It was ntoni van >eeuwenhoek that made it possible to imaŐe Ɵssue at a cellular level with the invenƟon of the liŐht microscope͘ The liŐht microscope is capable of imaŐinŐ Ɵssue, thouŐh eǆ vivo, with resoluƟons up to 1-2 ђm͘ However, absorpƟon and scaƩerinŐ of liŐht by Ɵssue hampers the imaŐinŐ depth͘ With the development of confocal microscopy, which uses point holes to eliminate out-of-focus liŐht, resoluƟon is improved to 0͘4 ђm͘ In contrary to CT, DRI and ultrasound, confocal microscopy has a limited imaŐinŐ depth, but hiŐh resoluƟon͘ When opƟcal coherence tomoŐraphy (OCT) was invented in 1991, the Őap between CT, DRI and ultrasound on one side of the depthͬresoluƟon-scale and confocal microscopy on the other side of that scale, was Įlled͘

&igƵre 1͘ Overview of most used imaŐinŐ technoloŐies͘ The associaƟon between depth resoluƟon (verƟcal

aǆis) and depth (horiǌontal aǆis) is shown͘ OCT is represented in oranŐe, ĮllinŐ the Őap between microscopy modaliƟes and the other techniƋues͘ hsed abbreviaƟons: (C)Dс(confocal) microscopy͖ OCTс opƟcal coherence tomoŐraphy͖ hSс ultrasound͖ SPECTс sinŐle photon emission computed tomoŐraphy͖ PETс positron emission tomoŐraphy͖ CTс computed tomoŐraphy͖ DRIс maŐneƟc resonance imaŐinŐ͘

(14)

Chapter 1 12

Besides ophthalmoloŐy, OCT is beinŐ eǆplored in other Įelds of specialƟes, such as cardioloŐy and uroloŐy ΀42-43΁͘

n imaŐe produced by OCT resembles Ɵssue architecture observed in histoloŐy and can therefore be considered as an ͚opƟcal biopsy͛͘ This ͚opƟcal biopsy͛ has hiŐh potenƟal in epithelial tumour diaŐnosis as it is non-invasive and real-Ɵme ΀38΁͘ Doreover, funcƟonal ƋuanƟtaƟve informaƟon can be eǆtracted, i͘e͘ Ňow informaƟon, layer thickness and aƩenuaƟon coeĸcient of the OCT-siŐnal ΀44-45΁͘ 'iven the match between OCT imaŐinŐ and histoloŐy in epithelial Ɵssues, OCT can play an important role in the diaŐnosis of (pre)maliŐnant lesions͘

acŬgroƵnĚ of techniƋƵe

OCT is based on depth resolved detecƟon of elasƟc liŐht scaƩerinŐ͘ When liŐht is directed at a Ɵssue sample, it will be parƟally back scaƩered͘ This back scaƩered liŐht is measured

OpƟcal coherence tomography

OCT is the opƟcal eƋuivalent of ultrasound, usinŐ liŐht instead of sound to produce imaŐes of Ɵssue͘ ResoluƟons up to 1-2 ђm can be achieved, beinŐ 100-250 Ɵmes hiŐher than hiŐh-resoluƟon ultrasound ΀38΁ and approachinŐ that of microscopy͘

In the mid 1980͛s the Įrst studies were published on one-dimensional, aǆial, opƟcal coherence tomoŐraphy (OCT)͘ These one-dimensional measurements were depth proĮle versus intensity scans and were called -scans, similar to ultrasound -scans ΀39-40΁͘ >ater, in 1991, cross-secƟonal or two-dimensional OCT was invented, i͘e͘ several -scans in a row created a two-dimensional B-scan, comparable to ultrasound B-scans (ĮŐure 2)͘ This cross-secƟonal imaŐinŐ is posiƟoned between ultrasound and (confocal) microscopy when it comes to the resoluƟonͬimaŐinŐ-depth Őap͘ Since 1991 OCT has been invesƟŐated thorouŐhly and scienƟĮc publicaƟons about OCT have increased eǆponenƟally (ĮŐure 3)͘ Especially in the Įeld of ophthalmoloŐy OCT has been eǆtensively studied (ĮŐure 4)͘ Ocular media is pracƟcally transparent which transmits liŐht with only minimal opƟcal aƩenuaƟon and scaƩerinŐ͘ Nowadays, OCT is used on a reŐular base in the ophthalmoloŐy clinic to diaŐnose corneal diseases and reƟnal diseases, like Őlaucoma ΀41΁͘

&igƵre 2͘ Build-up of a dataset of the skin͘ On the leŌ a 1-D aǆial scan, showinŐ reŇecƟvity vs͘ depth͘ In the

middle a cross-secƟonal imaŐe or B-scan composed of consecuƟve -scans in transverse direcƟon͘ On the riŐht a 3-D imaŐe made of consecuƟve lonŐitudinal B-scans͘

(15)

1

&igƵre 3͘ ll peer-reviewed publicaƟons in OCT to date (2014 Thomson Reuters)͘ On the verƟcal aǆis the number

of publicaƟons͖ on the horiǌontal aǆis the year of publicaƟon͘

&igƵre 4͘ ll peer-reviewed publicaƟons on OCT cateŐoriǌed accordinŐ to the diīerent Įelds in science where

(16)

Chapter 1 14

at diīerent depths at a parƟcular locaƟon on the Ɵssue usinŐ low-coherence interferometry resulƟnŐ in a reŇecƟon proĮle in the depth (ǌ-) direcƟon͘ The maŐnitude of the OCT siŐnal at each depth is determined by the diīerent cellular structures in the imaŐed volume (typically in the order of 103_ʅm3_{or 1 picoliter) and as a result it diīers per Ɵssue type͘ Several adjacent}

depth proĮles can be acƋuired in the lateral (ǆ-) direcƟon and displayed as a Őrey-scale imaŐe in real Ɵme which is known as an OCT B-scan͘ SubseƋuently, the OCT beam can be scanned across a Ɵssue sample in the other lateral (y-) direcƟon, resulƟnŐ in a 3-dimensional imaŐe representaƟon with acƋuisiƟon speed reported up to several volumes per second͘

System setƵp

The OCT system (ĮŐure 5) based on the Dichelson interferometer, consists of Įber-opƟc components, illuminated by a broad wavelenŐth-ranŐe liŐht source operaƟnŐ in the near-infrared (typically 1250 nm ʹ 1350 nm for non-ophthalmic applicaƟons)͘

&igƵre 5͘ SchemaƟc overview of an OCT system based on the Dichelson interferometer (which consist of a liŐht

source, a liŐht spliƩer and a detectorͬspectrometer)͘ n opƟcal beam (from the liŐht source) is split into two arms͘ One arm is directed at the Ɵssue (scanninŐ sample arm), the other at a mirror (the reference arm)͘ The reŇected liŐht from these two paths is recombined and the diīerences between these two paths can be shown in a Őreyscale imaŐe͘

small fracƟon of the liŐht is Őuided towards a ͚reference͛ mirror͖ the majority is directed to the Ɵssue, usinŐ handheld yz-scanninŐ devices or miniaturiǌed endoscopic probes (ĮŐure 6)͘ The handheld scanninŐ device is held aŐainst the Ɵssue to obtain the OCT-imaŐes͘ Both fracƟons are combined and directed towards a detecƟon unit and are processed on a computer͘ Disposable in-vivo OCT-probes with a diameter less than 1 mm are developed and already used in several medical seƫnŐs, enablinŐ OCT to be used in endoscopes or in combinaƟon with

(17)

1

16'ͬ18'-needles to access internal Ɵssue͘ Clinical value of OCT-imaŐes depends on obvious factors such as hiŐh resoluƟon, hiŐh imaŐinŐ speed and adeƋuate contrast to discriminate between beniŐn and maliŐnant Ɵssues͘

&igƵre 6͘ Clinical OCT system employed at our medical center͘ >eŌ shows the system on a cart allowinŐ the

system to be transferred from diīerent clinical seƫnŐs easily͘ The system can be interfaced with either a ǆͬy scanninŐ handheld device (upper riŐht) or a rotaƟnŐ endoscopic scanninŐ device (lower riŐht)͘ The endoscopic device scans its environment by fast rotaƟonal (liŐhthouse like) movements͘ 3D imaŐinŐ in this case is achieved by a pullback of the probe͘

YƵalitaƟǀe anĚ ƋƵanƟtaƟǀe OCT analysis

Dost research on OCT and epithelial cancers focuses on ƋualitaƟve imaŐinŐ, i͘e͘ on morpholoŐic aspects, and correlaƟon of these aspects with histopatholoŐy͘ One of these ƋualitaƟve, morpholoŐic aspects that is studied freƋuently in OCT imaŐes, is Ɵssue architecture͘ For eǆample, healthy skin shows a clear layered skin architecture in OCT imaŐes, while basal cell carcinoma (BCC) shows loss of layers and disarranŐement of the layered architecture ΀46-48΁͘ Several other studies have shown that OCT is capable of diīerenƟaƟnŐ layers in epithelial Ɵssue, such as in the larynǆ ΀49΁, esophaŐus ΀50΁, bladder ΀51-52΁ and in cervical epithelial Ɵssue ΀53΁͘ Doreover, irreŐulariƟes of the epithelial layer and disrupƟon of the basal membrane could also be detected with OCT imaŐinŐ of vulvar skin of paƟents with PaŐet͛s disease ΀53΁͘ Other studies have focused on details within the OCT imaŐes, for eǆample verƟcal

(18)

Chapter 1 16

cone-shaped structures were found in OCT imaŐes of maliŐnant melanoma ΀54΁͘ Furthermore, comparison of the OCT imaŐes with histopatholoŐy has been invesƟŐated in non-melanoma skin cancer ΀47΁, in piŐmented melanoctyic lesions ΀55΁, in oral sƋuamous cell carcinoma ΀56΁ and in the esophaŐus and esophaŐeal dysplasia ΀57-60΁͘ In summary, OCT is capable of producinŐ real-Ɵme hiŐh resoluƟon imaŐes showinŐ Ɵssue structure and basement membrane inteŐrity comparable with histopatholoŐy͘

Besides ƋualitaƟve imaŐinŐ, OCT holds the potenƟal to provide ƋuanƟtaƟve informaƟon about the imaŐed Ɵssue as well͘ This can be achieved by advanced analysis of the OCT siŐnal throuŐh calculaƟnŐ the opƟcal aƩenuaƟon coeĸcient with Beer͛s law͘ The aƩenuaƟon coeĸcient ʅ_OCT ΀mm-1_{΁ describes the decay of the measured OCT siŐnal, which is correlated to the architecture}

and content of the measured Ɵssue (ĮŐure 7)͘

&igƵre 7͘ (a) Three-dimensional (3-D) representaƟon of 15 by 15 by 3 mm OCT scan͖ (b) two-dimensional

(2-D) cross-secƟonal imaŐe with the reŐion of interest (ROI) depicted in red͘ The epithelial layer is shown as the second dark Őray layer in the cross-secƟonal imaŐe͖ (c) averaŐe -scan obtained from the ROI in the 2-D OCT scan͘ The thickness of the epithelial layer is measured in this Őraph and is represented as d͘ ƩenuaƟon Įt (ʅ_OCT) is represented by the slope of the OCT siŐnal shown in transparent red͘

(19)

1

The aƩenuaƟon coeĸcient is the result of absorpƟon and scaƩerinŐ of liŐht͘ Chromophores in the measured Ɵssue determine the absorpƟon of liŐht, whereas siǌe and refracƟve indeǆ of the measured content determine the scaƩerinŐ of the liŐht͘ >iŐht scaƩerinŐ measurements are sensiƟve to variaƟons in Ɵssue morpholoŐy (density) at sub-wavelenŐth scales (i͘e͘ у 650 nm) ΀61΁͘ When cells boost their DN durinŐ the cell cycle ΀62΁, the refracƟve indeǆ of the nucleus, and hence the liŐht scaƩerinŐ property, increases͘ It is known that cancerous Ɵssue has a hiŐh proporƟon of dividinŐ cells ΀63΁ which increases their DN fracƟon͘ In dysplasƟc cells, like the cells in VIN, DN replicaƟon takes place too ΀64΁ and chanŐes in scaƩerinŐ properƟes can be eǆpected as well͘ Our ƋuanƟtaƟve OCT measurements are sensiƟve to chanŐes on lenŐth scales of around ࡥͬ2 у 650 nm, e͘Ő͘ on the scale of orŐanelles and cells ΀65΁͘ Thus, diīerence in aƩenuaƟon coeĸcient in diīerent Ɵssue-types miŐht be result of the amount of dividinŐ cells and conseƋuently the chanŐe in scaƩerinŐ properƟes͘

In conclusion, as Ɵssue content chanŐes when maliŐnancies develop, OCT may have the potenƟal as monitorinŐ and diaŐnosinŐ tool as it allows ƋuanƟtaƟve analysis of scaƩerinŐ properƟes which may alter durinŐ cancer development͘

THESIS MOTIVATION AND AIM

OCT imaŐinŐ has a Őreat potenƟal in servinŐ as an ͚opƟcal biopsy͛ tool because OCT imaŐes resemble Ɵssue architecture similar to histopatholoŐy͘ Doreover, the aƩenuaƟon coeĸcient of the OCT siŐnal can be determined͖ enablinŐ not only ƋualitaƟve, but ƋuanƟtaƟve analysis as well͘

In this thesis we used a commercially available OCT system operaƟnŐ at 1300 nm to invesƟŐate diīerent epithelial (pre)maliŐnancies͘ The aim of the research is to present the potenƟal of OCT imaŐinŐ in discriminaƟnŐ normal Ɵssue from (pre)maliŐnant Ɵssue in the ŐynecoloŐical, uroloŐical and dermatoloŐical clinic͘ The OCT imaŐes were analyǌed ƋualitaƟvely on morpholoŐy and ƋuanƟtaƟvely by usinŐ the aƩenuaƟon coeĸcient͘ Furthermore, a learninŐ study was conducted to invesƟŐate inter-observer diīerences and the learninŐ-curve in OCT imaŐinŐ analysis͘

THESIS STRUCTURE

summary of the principles of OCT and an overview of the current literature on the diaŐnosƟc value of OCT in the diaŐnosis of epithelial (pre)maliŐnant lesions is Őiven in chapter 2͘ Chapter 3 shows the possibility of OCT imaŐinŐ in diīerenƟaƟnŐ normal vulvar Ɵssue from vulvar intraepithelial neoplasia (VIN)͘ Because these Őenital premaliŐnant lesions can develop into maliŐnant carcinomas that have to be eǆcised, the potenƟal of OCT in determininŐ resecƟon

(20)

Chapter 1 18

marŐins durinŐ surŐery of vulvar carcinoma is studied in chapter 4͘ Similar to VIN in chapter 3, the discriminaƟnŐ potenƟal of OCT in penile intraepithelial neoplasia is studied in chapter 5͘ In chapter 6, OCT is used to imaŐe piŐmented lesions of the skin and to invesƟŐate whether OCT can discriminate between beniŐn and maliŐnant piŐmented skin lesions͘ s determinaƟon of the aƩenuaƟon coeĸcient of the OCT siŐnal is observer-dependent, we studied the inter-observer variability and the learninŐ curve of ƋuanƟtaƟve OCT analysis in chapter 7͘ In chapter 8 our conclusions and future perspecƟves are presented͘

(21)

1 REFERENCES

΀1΁ SieŐel R, Naishadham D, :emal ͘ Cancer staƟsƟcs, 2012͘ C Cancer : Clin 2012͖ 62(1): 10-29͘ ΀2΁ DcConnell TH͘ the Nature of Disease͘ >ippincoƩ Williams Θ Wilkinson͘ 2007͘

΀3΁ :udson P>, Habermann EB, Baǆter NN, Durham SB, VirniŐ B͘ Trends in the incidence of invasive and in situ vulvar carcinoma͘ Obstet 'ynecol 2006͖ 107(5): 1018-1022͘

΀4΁ van de Nieuwenhof HP, DassuŐer >F, van der voort I, Bekkers R>, Casparie D, bma W, et al͘ Vulvar sƋuamous cell carcinoma development aŌer diaŐnosis of VIN increases with aŐe͘ Eur : Cancer 2009͖ 45(5): 851-856͘

΀5΁ van der voort I, ShiranŐo H, Hoevenaars BD, 'reŌe :D, de Hullu :, de Wilde PC, et al͘ Vulvar sƋuamous cell carcinoma is a mulƟfactorial disease followinŐ two separate and independent pathways͘ Int : 'ynecol Pathol 2006͖ 25(1): 22-29͘

΀6΁ PreƟ D, Van Seters D, Sideri D, van Beurden D͘ SƋuamous vulvar intraepithelial neoplasia͘ Clin Obstet 'ynecol 2005͖ 48(4): 845-861͘

΀7΁ nastasiadis P, Skaphida P, <outlaki N, Boli , 'alaǌios ', >iberis V͘ Trends in epidemioloŐy of preinvasive and invasive vulvar neoplasias͘ 13 year retrospecƟve analysis in Thrace, 'reece͘ rch 'ynecol Obstet 2000͖ 264(2): 74-79͘

΀8΁ Iversen T, Tretli S͘ Intraepithelial and invasive sƋuamous cell neoplasia of the vulva: trends in incidence, recurrence, and survival rate in Norway͘ Obstet 'ynecol 1998͖ 91(6): 969-972͘

΀9΁ ndersen B>, TurnƋuist D, >aPolla :, Turner D͘ Seǆual funcƟoninŐ aŌer treatment of in situ vulvar cancer: preliminary report͘ Obstet 'ynecol 1988͖ 71(1): 15-19͘

΀10΁ Van Seters D, van Beurden D, de Craen :͘ Is the assumed natural history of vulvar intraepithelial neoplasia III based on enouŐh evidence͍ systemaƟc review of 3322 published paƟents͘ 'ynecol Oncol 2005͖ 97(2): 645-651͘

΀11΁ Wills , Obermair ͘ review of complicaƟons associated with the surŐical treatment of vulvar cancer͘ 'ynecol Oncol 2013͖ 131(2): 467-479͘

΀12΁ Chan :<, SuŐiyama V, Pham H, 'u D, RutŐers :, Osann <, et al͘ DarŐin distance and other clinico-patholoŐic proŐnosƟc factors in vulvar carcinoma: a mulƟvariate analysis͘ 'ynecol Oncol 2007͖ 104(3): 636-641͘ ΀13΁ Heaps :D, Fu zS, Dontǌ F:, Hacker NF, Berek :S͘ SurŐical-patholoŐic variables predicƟve of local recurrence

in sƋuamous cell carcinoma of the vulva͘ 'ynecol Oncol 1990͖ 38(3): 309-314͘

΀14΁ De Hullu :, Hollema H, >olkema S, Boeǌen D, Boonstra H, BurŐer DP, et al͘ Vulvar carcinoma͘ The price of less radical surŐery͘ Cancer 2002͖ 95(11): 2331-2338͘

΀15΁ ndersen B>, Hacker NF͘ Psychoseǆual adjustment aŌer vulvar surŐery͘ Obstet 'ynecol 1983͖ 62(4): 457-462͘

΀16΁ Rubin D, <leter B, hou D, yala ', Cubilla >, Yuint W', et al͘ DetecƟon and typinŐ of human papillomavirus DN in penile carcinoma: evidence for mulƟple independent pathways of penile carcinoŐenesis͘ m : Pathol 2001͖ 159(4): 1211-1218͘

΀17΁ WabinŐa HR, Parkin DD, Wabwire-DanŐen F, Nambooǌe S͘ Trends in cancer incidence in <yadondo County, hŐanda, 1960-1997͘ Br : Cancer 2000͖ 82(9): 1585-1592͘

΀18΁ Parkin DD, Duir CS͘ Cancer Incidence in Five ConƟnents͘ Comparability and Ƌuality of data͘ IRC Sci Publ 1992͖ 120: 45-173͘

΀19΁ Daiche '͘ EpidemioloŐical aspects of cancer of the penis in Finland͘ Eur : Cancer Prev 1992͖ 1(2): 153-158͘

΀20΁ Cubilla >, Deijer C:, zounŐ RH͘ DorpholoŐical features of epithelial abnormaliƟes and precancerous lesions of the penis͘ Scand : hrol Nephrol Suppl 2000͖ 205: 215-219͘

΀21΁ Wikstrom , Hedblad D, Syrjanen S͘ Penile intraepithelial neoplasia: histopatholoŐical evaluaƟon, HPV typinŐ, clinical presentaƟon and treatment͘ : Eur cad Dermatol Venereol 2012͖ 26(3): 325-330͘

΀22΁ von <roŐh ', Horenblas S͘ The manaŐement and prevenƟon of premaliŐnant penile lesions͘ Scand : hrol Nephrol Suppl 2000͖ 205: 220-229͘

(22)

Chapter 1 20

΀23΁ Dadan V, >ear :T, Sǌeimies RD͘ Non-melanoma skin cancer͘ >ancet 2010͖ 375(9715): 673-685͘

΀24΁ Neville :, Welch E, >eīell D:͘ DanaŐement of nonmelanoma skin cancer in 2007͘ Nat Clin Pract Oncol 2007͖ 4(8): 462-469͘

΀25΁ InternaƟonal Őency for Research on Cancer͘ '>OBOCN 2012 EsƟmated Cancer Incidence, Dortality and Prevalence worldwide in 2012͘ ΀WWW document΁ hR> hƩp:ͬͬŐlobocan͘iarc͘frͬPaŐesͬfactͺsheetsͺ populaƟon͘aspǆ (last accessed: 14 Darch 2014)͘

΀26΁ 'eller C, Halpern C͘ The ever-evolvinŐ landscape for detecƟon of early melanoma: challenŐes and promises͘ : Invest Dermatol 2013͖ 133(3): 583-585͘

΀27΁ <iƩler H, PehamberŐer H, Wolī <, Binder D͘ DiaŐnosƟc accuracy of dermoscopy͘ >ancet Oncol 2002͖ 3(3): 159-165͘

΀28΁ Campos-do-Carmo ', Ramos-e-Silva D͘ Dermoscopy: basic concepts͘ Int : Dermatol 2008͖ 47(7): 712-719͘ ΀29΁ nnessi ', Bono R, SampoŐna F, FaraŐŐiana T, beni D͘ SensiƟvity, speciĮcity, and diaŐnosƟc accuracy of

three dermoscopic alŐorithmic methods in the diaŐnosis of doubƞul melanocyƟc lesions: the importance of liŐht brown structureless areas in diīerenƟaƟnŐ atypical melanocyƟc nevi from thin melanomas͘ : m cad Dermatol 2007͖ 56(5): 759-767͘

΀30΁ Carrera C, PuiŐ S, Dalvehy :͘ In vivo confocal reŇectance microscopy in melanoma͘ Dermatol Ther 2012͖ 25(5): 410-422͘

΀31΁ Rajadhyaksha D, 'rossman D, Esterowitǌ D, Webb RH, nderson RR͘ In vivo confocal scanninŐ laser microscopy of human skin: melanin provides stronŐ contrast͘ : Invest Dermatol 1995͖ 104(6): 946-952͘ ΀32΁ 'uitera P, Pellacani ', >onŐo C, Seidenari S, vramidis D, Denǌies SW͘ In vivo reŇectance confocal microscopy

enhances secondary evaluaƟon of melanocyƟc lesions͘ : Invest Dermatol 2009͖ 129(1): 131-138͘

΀33΁ Pellacani ', 'uitera P, >onŐo C, vramidis D, Seidenari S, Denǌies S͘ The impact of in vivo reŇectance confocal microscopy for the diaŐnosƟc accuracy of melanoma and eƋuivocal melanocyƟc lesions͘ : Invest Dermatol 2007͖ 127(12): 2759-2765͘

΀34΁ SeŐura S, PuiŐ S, Carrera C, Palou :, Dalvehy :͘ Development of a two-step method for the diaŐnosis of melanoma by reŇectance confocal microscopy͘ : m cad Dermatol 2009͖ 61(2): 216-229͘

΀35΁ Pellacani ', Cesinaro D, Seidenari S͘ ReŇectance-mode confocal microscopy for the in vivo characteriǌaƟon of paŐetoid melanocytosis in melanomas and nevi͘ : Invest Dermatol 2005͖ 125(3): 532-537͘

΀36΁ >anŐley R', Rajadhyaksha D, Dwyer P:, Sober :, FloƩe T:, nderson RR͘ Confocal scanninŐ laser microscopy of beniŐn and maliŐnant melanocyƟc skin lesions in vivo͘ : m cad Dermatol 2001͖ 45(3): 365-376͘ ΀37΁ 'erŐer , <oller S, <ern T, Dassone C, SteiŐer <, RichƟŐ E, et al͘ DiaŐnosƟc applicability of in vivo confocal

laser scanninŐ microscopy in melanocyƟc skin tumours͘ : Invest Dermatol 2005͖ 124(3): 493-498͘ ΀38΁ Dreǆler W, Fujimoto : (eds)͘ OpƟcal Coherence TomoŐraphy: TechnoloŐy and pplicaƟons 2008th edn͘

SprinŐer, New zork͘

΀39΁ Fercher F, DenŐedoht <, Werner W͘ Eye-lenŐth measurement by interferometry with parƟally coherent liŐht͘ Opt >eƩ 1988͖ 13(3): 186-188͘

΀40΁ Fujimoto :', De Silvestri S, Ippen EP, PuliaĮto C, DarŐolis R, Oseroī ͘ Femtosecond opƟcal ranŐinŐ in bioloŐical systems͘ Opt >eƩ 1986͖ 11(3): 150͘

΀41΁ Hee DR, IǌaƩ :, Swanson E, HuanŐ D, Schuman :S, >in CP, et al͘ OpƟcal coherence tomoŐraphy of the human reƟna͘ rch Ophthalmol 1995͖ 113(3): 325-332͘

΀42΁ zun SH, Tearney ':, Vakoc B:, Shishkov D, Oh Wz, Desjardins E, et al͘ Comprehensive volumetric opƟcal microscopy in vivo͘ Nat Ded 2006͖ 12(12): 1429-1433͘

΀43΁ Barwari <, de Bruin DD, Faber D:, van >eeuwen T', de la RoseƩe ::, >aŐuna DP͘ DiīerenƟaƟon between normal renal Ɵssue and renal tumours usinŐ funcƟonal opƟcal coherence tomoŐraphy: a phase I in vivo human study͘ B:h Int 2012͖ 110(8PtB): E415-E420͘͘

΀44΁ van eeburŐ E:, Cereda D', van der Schoot :, PerƟle ', van Deurs :C͘ Early perfusion of a free RPE-choroid ŐraŌ in paƟents with eǆudaƟve macular deŐeneraƟon can be imaŐed with spectral domain-OCT͘ Invest Ophthalmol Vis Sci 2011͖ 52(8): 5881-5886͘

΀45΁ Srinivasan V:, Radhakrishnan H, >o EH, Dandeville ET, :ianŐ :z, Barry S, et al͘ OCT methods for capillary velocimetry͘ Biomed Opt Eǆpress 2012͖ 3(3): 612-629͘

(23)

1

΀46΁ 'ambichler T, Orlikov , Vasa R, Doussa ', Hoīmann <, Stucker D, et al͘ In vivo opƟcal coherence

tomoŐraphy of basal cell carcinoma͘ : Dermatol Sci 2007͖ 45(3): 167-173͘

΀47΁ DoŐensen D, :oerŐensen TD, NurnberŐ BD, Dorsy H, Thomsen :B, Thrane >, et al͘ ssessment of opƟcal coherence tomoŐraphy imaŐinŐ in the diaŐnosis of non-melanoma skin cancer and beniŐn lesions versus normal skin: observer-blinded evaluaƟon by dermatoloŐists and patholoŐists͘ Dermatol SurŐ 2009͖ 35(6): 965-972͘ ΀48΁ Strasswimmer :, Pierce DC, Park BH, Neel V, de Boer :F͘ PolariǌaƟon-sensiƟve opƟcal coherence tomoŐraphy

of invasive basal cell carcinoma͘ : Biomed Opt 2004͖ 9(2): 292-298͘

΀49΁ WonŐ B:, :ackson RP, 'uo S, RidŐway :D, Dahmood h, Su :, et al͘ In vivo opƟcal coherence tomoŐraphy of the human larynǆ: normaƟve and beniŐn patholoŐy in 82 paƟents͘ >arynŐoscope 2005͖ 115(11): 1904-1911͘ ΀50΁ Cilesiǌ I, Fockens P, <erindonŐo R, Faber D, TytŐat ', Ten <ate F, et al͘ ComparaƟve opƟcal coherence

tomoŐraphy imaŐinŐ of human esophaŐus: how accurate is localiǌaƟon of the muscularis mucosae͍ 'astrointest Endosc 2002͖ 56(6): 852-857͘

΀51΁ :esser C, Boppart S, Pitris C, Stamper D>, Nielsen 'P, Breǌinski DE, et al͘ HiŐh resoluƟon imaŐinŐ of transiƟonal cell carcinoma with opƟcal coherence tomoŐraphy: feasibility for the evaluaƟon of bladder patholoŐy͘ Br : Radiol 1999͖ 72(864): 1170-1176͘

΀52΁ Tearney ':, Breǌinski DE, Southern :F, Bouma BE, Boppart S, Fujimoto :'͘ OpƟcal biopsy in human uroloŐic Ɵssue usinŐ opƟcal coherence tomoŐraphy͘ : hrol 1997͖ 157(5): 1915-1919͘

΀53΁ Escobar PF, Belinson :>, White , Shakhova ND, Feldchtein FI, <areta DV, et al͘ DiaŐnosƟc eĸcacy of opƟcal coherence tomoŐraphy in the manaŐement of preinvasive and invasive cancer of uterine cerviǆ and vulva͘ Int : 'ynecol Cancer 2004͖ 14(3): 470-474͘

΀54΁ 'ambichler T, ReŐeniter P, Bechara F', Orlikov , Vasa R, Doussa ', et al͘ CharacteriǌaƟon of beniŐn and maliŐnant melanocyƟc skin lesions usinŐ opƟcal coherence tomoŐraphy in vivo͘ : m cad Dermatol 2007͖ 57(4): 629-637͘

΀55΁ de 'iorŐi V, Stante D, Dassi D, Davilia >, CappuŐi P, Carli P͘ Possible histopatholoŐic correlates of dermoscopic features in piŐmented melanocyƟc lesions idenƟĮed by means of opƟcal coherence tomoŐraphy͘ Eǆp Dermatol 2005͖ 14(1): 56-59͘

΀56΁ Wilder-Smith P, >ee <, 'uo S, hanŐ :, Osann <, Chen , et al͘ In vivo diaŐnosis of oral dysplasia and maliŐnancy usinŐ opƟcal coherence tomoŐraphy: preliminary studies in 50 paƟents͘ >asers SurŐ Ded 2009͖ 41(5): 353-357͘

΀57΁ Evans :, Poneros :D, Bouma BE, Bressner :, Halpern EF, Shishkov D, et al͘ OpƟcal coherence tomoŐraphy to idenƟfy intramucosal carcinoma and hiŐh-Őrade dysplasia in BarreƩ͛s esophaŐus͘ Clin 'astroenterol Hepatol 2006͖ 4(1): 38-43͘

΀58΁ IsenberŐ ', Sivak DV, :r͘, Chak , WonŐ RC, Willis :E, Wolf B, et al͘ ccuracy of endoscopic opƟcal coherence tomoŐraphy in the detecƟon of dysplasia in BarreƩ͛s esophaŐus: a prospecƟve, double-blinded study͘ 'astrointest Endosc 2005͖ 62(6): 825-831͘

΀59΁ Poneros :D, Brand S, Bouma BE, Tearney ':, Compton CC, Nishioka NS͘ DiaŐnosis of specialiǌed intesƟnal metaplasia by opƟcal coherence tomoŐraphy͘ 'astroenteroloŐy 2001͖ 120(1): 7-12͘

΀60΁ uccaro ', 'ladkova N, VarŐo :, Feldchtein F, aŐaynova E, Conwell D, et al͘ OpƟcal coherence tomoŐraphy of the esophaŐus and proǆimal stomach in health and disease͘ m : 'astroenterol 2001͖ 96(9): 2633-2639͘ ΀61΁ Duller D', Valdeǌ T, 'eorŐakoudi I, Backman V, Fuentes C, <abani S, et al͘ Spectroscopic detecƟon and

evaluaƟon of morpholoŐic and biochemical chanŐes in early human oral carcinoma͘ Cancer 2003͖ 97(7): 1681-1692͘

΀62΁ Bista R<, hƩam S, WanŐ P, Staton <, Choi S, Bakkenist C:, et al͘ YuanƟĮcaƟon of nanoscale nuclear refracƟve indeǆ chanŐes durinŐ the cell cycle͘ : Biomed Opt 2011͖ 16(7): 070503͘

΀63΁ Williams 'H, Stoeber <͘ The cell cycle and cancer͘ : Pathol 2012͖ 226(2): 352-364͘

΀64΁ Freeman , Dorris >S, Dills D, Stoeber <, >askey R, Williams 'H, et al͘ Dinichromosome maintenance proteins as bioloŐical markers of dysplasia and maliŐnancy͘ Clin Cancer Res 1999͖ 5(8): 2121-2132͘ ΀65΁ van der Deer F:, Faber D:, alders DC, Poot , Vermes I, van >eeuwen T'͘ poptosis- and

necrosis-induced chanŐes in liŐht aƩenuaƟon measured by opƟcal coherence tomoŐraphy͘ >asers Ded Sci 2010͖ 25(2): 259-267͘

(24)

(25)

R͘ Wessels D͘D͘ de Bruin D͘:͘ Faber T͘'͘ van >eeuwen D͘ van Beurden T͘:͘ Ruers

R͘ Wessels and D͘D͘ de Bruin contributed eƋually to this paper͘

OpƟcal biopsy of epithelial cancers by opƟcal

coherence tomoŐraphy (OCT)

Lasers Med Sci 2014; 29(3): 1297-305

(26)

Chapter 2 24

ABSTRACT

OpƟcal coherence tomoŐraphy (OCT) is an opƟcal techniƋue that measures the backscaƩerinŐ of near-infrared liŐht by Ɵssue͘ OCT yields in 2D and 3D imaŐes at micrometer-scale resoluƟon, thus providinŐ opƟcal biopsies, approachinŐ the resoluƟon of histopatholoŐical imaŐinŐ͘ The techniƋue has shown to allow in vivo diīerenƟaƟon between beniŐn and maliŐnant epithelial Ɵssue, throuŐh ƋualitaƟve assessment of OCT-imaŐes, as well as by ƋuanƟtaƟve evaluaƟon e͘Ő͘ funcƟonal OCT͘ This study aims to summariǌe the principles of OCT and to discuss the current literature on the diaŐnosƟc value of OCT in the diaŐnosis of epithelial (pre)maliŐnant lesions͘ The authors did a systemaƟc search of the electronic databases Pubmed and Embase on OCT in the diaŐnosƟc process of (pre)maliŐnant epithelial lesions͘ OCT is able to diīerenƟate between beniŐn and (pre)maliŐnant lesions of epithelial oriŐin in a wide variety of Ɵssues͘ In this way, OCT can detect skin cancers, oral, larynŐeal and esophaŐeal cancer as well as Őenital and bladder cancer͘ OCT is an innovaƟve techniƋue which enables an opƟcal biopsy of epithelial lesions͘ The incorporaƟon of OCT in speciĮc tools, like handheld and catheter-based probes, will further improve to the implementaƟon of this technoloŐy in daily clinical pracƟce͘

(27)

OpƟcal biopsy of epithelial cancers by opƟcal coherence tomoŐraphy (OCT) 25

2 INTRODUCTION

Dore than 80й of all cancers oriŐinate from the epithelium ΀1΁͘ Current methods of diaŐnosinŐ these cancers rely on histoloŐical and cytoloŐical eǆaminaƟon of Ɵssue or body Ňuids͘ For this purpose, brushes or biopsies are harvested, which have to be Įǆed and stained before diaŐnosis͘ To circumvent this Ɵme intensive procedure, new opƟcal imaŐinŐ modaliƟes that enable real Ɵme (pre)maliŐnant lesions detecƟon in vivo are beinŐ invesƟŐated at the moment͘ OpƟcal coherence tomoŐraphy (OCT) is a technoloŐy developed in the early 1990s for ophthalmoloŐical applicaƟons ΀2-3΁ and is profoundly used in that seƫnŐ͘ OCT is the opƟcal eƋuivalent of ultrasound, usinŐ liŐht instead of sound to produce imaŐes of Ɵssue͘ ResoluƟons up to 1-2 ђm can be achieved, beinŐ 100-250 Ɵmes hiŐher than hiŐh-resoluƟon ultrasound ΀4΁ and approachinŐ that of microscopy͘ However, due to liŐht scaƩerinŐ by the sample, imaŐinŐ depth is usually limited to Ε2 mm͘ n imaŐe produced by OCT resembles the Ɵssue architecture observed in histoloŐy and can therefore be considered as an ͞opƟcal biopsy͘͟ This ͞opƟcal biopsy͟ has hiŐh potenƟal in epithelial tumor diaŐnosis as it is non-invasive and real-Ɵme ΀4΁͘ Doreover, funcƟonal ƋuanƟtaƟve informaƟon can be eǆtracted, i͘e͘ Ňow informaƟon, layer thickness and aƩenuaƟon coeĸcient of the OCT-siŐnal ΀5-6΁͘ 'iven the match between OCT imaŐinŐ and histoloŐy in epithelial Ɵssues, OCT can play an important role in the diaŐnosis of tumorous lesions͘

Ōer our pioneerinŐ work in kidney cancer ΀7-8΁, bladder cancer ΀9΁ and premaliŐnant vulvar intraepithelial neoplasia ΀10΁, we aim to review the diaŐnosƟc potenƟal of OCT in (pre) cancerous lesions of the human epithelium͘ s most invesƟŐators study ƋualitaƟve OCT, we focus on ƋualitaƟve, morpholoŐical imaŐinŐ of human epithelium, especially of the skin, the oropharynŐeal and larynŐeal area, the esophaŐus, the Őenital area, and the bladder͘ t the end, we focus on ƋuanƟtaƟve, e͘Ő͘ funcƟonal OCT and we outline the potenƟal future applicaƟon of funcƟonal OCT imaŐinŐ in diaŐnosinŐ (pre)cancer͘

BacŬgroƵnĚ of techniƋƵe

OCT is based on depth resolved detecƟon of elasƟc liŐht scaƩerinŐ͘ When liŐht is directed at a Ɵssue sample, it will be parƟally back scaƩered͘ This back-scaƩered liŐht is measured at diīerent depths at a parƟcular locaƟon on the Ɵssue usinŐ low-coherence interferometry resulƟnŐ in a reŇecƟon proĮle in the depth (z-) direcƟon͘ The maŐnitude of the OCT siŐnal at each depth is determined by the diīerent cellular structures in the imaŐed volume (typically in the order of 103 _ʅm3_{or 1 pl) and as a result it diīers per Ɵssue type͘ Several adjacent depth}

proĮles can be acƋuired in the lateral (ǆ-) direcƟon and displayed as a Őrey-scale imaŐe in real Ɵme which is known as an OCT B-scan͘ SubseƋuently, the OCT beam can be scanned across a Ɵssue sample in the other lateral (y-) direcƟon, resulƟnŐ in a 3D imaŐe representaƟon with acƋuisiƟon speed reported up to several volumes per second͘

(28)

Chapter 2 26

System setƵp

The OCT system (FiŐ͘ 1) consists of an interferometer, Őenerally constructed from Įber opƟc components, illuminated by a broad wavelenŐth-ranŐe liŐht source operaƟnŐ in the near-infrared (typically 1250 nm ʹ 1350 nm for non-ophthalmic applicaƟons)͘ small fracƟon of the liŐht is Őuided towards a ͞reference͟ mirror; the majority is directed to the Ɵssue, usinŐ handheld yz-scanninŐ devices or miniaturiǌed endoscopic probes (FiŐ͘ 2)͘ Both fracƟons are combined and directed towards a detecƟon unit and subseƋuent computer processinŐ͘ Disposable in vivo OCT-probes with a diameter less than 1 mm are developed and already used in several medical seƫnŐs, enablinŐ OCT to be used in endoscopes or in combinaƟon with 16'ͬ18' needles to access internal Ɵssue͘ Clinical value of OCT-imaŐes depends on obvious factors such as hiŐh resoluƟon, hiŐh-imaŐinŐ speed and adeƋuate contrast to discriminate between beniŐn and maliŐnant Ɵssues͘

Contrast in OCT is caused by spaƟal diīerences in refracƟve indeǆ of diīerent Ɵssue consƟtuents, e͘Ő͘, contrast oriŐinates from reŇecƟon of diīerent structures͘ It is known that the refracƟve indeǆ is proporƟonal to the density of the cells and cell structure͘ Because maliŐnant cells display an increased number, larŐer and more irreŐularly shaped nuclei with a hiŐher refracƟve indeǆ and more acƟve mitochondria, OCT imaŐes are eǆpected to be diīerent in maliŐnant Ɵssue compared to normal and beniŐn Ɵssue ΀11΁͘

FigƵre 1͘ SchemaƟc overview of the OCT system͘ n opƟcal beam (from the liŐht source) is split into two arms͘

One arm is directed at the Ɵssue (scanninŐ sample arm), the other at a mirror (the reference arm)͘ The reŇected liŐht from these two paths is recombined and the diīerences between these two paths can be shown in a

(29)

2 METHODS

For this systemaƟc review, we performed a literature search usinŐ PubDed and Embase for oriŐinal and review arƟcles wriƩen in EnŐlish, French or 'erman usinŐ the search terms oƉƟcal

coherence tomography and skin cancer or melanoma (renderinŐ 85 and 16 articles, respectively),

opƟcal coherence tomography and oral͕ pharyngeal͕ or laryngeal cancer ;renderinŐ 30 and 8 arƟcles respecƟvely), opƟcal coherence tomography and esophageal diseases (renderinŐ 74 and 27 arƟcles, respecƟvely)͘ OpƟcal coherence tomography and Ƶrinary ďladder͕ ǀƵlǀar͕ cerǀical or

penile cancer (renderinŐ 95 and 42 arƟcles, respecƟvely)͘ This search resulted in a total of 377

hits, from which we selected 57 arƟcles based on relevant contribuƟon in describinŐ technoloŐy and clinical evaluaƟon of these methods in skin cancer, oral, larynŐeal and esophaŐeal cancer and Őenital and bladder cancer͘ Selected arƟcles oriŐinated from 1991 to 2012͘

FigƵre 2͘ Clinical OCT system employed at our medical center͘ LeŌ shows the system on a cart allowinŐ the

system to be transferred from diīerent clinical seƫnŐs easily͘ The system can be interfaced with either an ǆͬy scanninŐ handheld device (Ƶpper right) or a rotaƟnŐ endoscopic scanninŐ device (lower right)͘ The laƩer scans its environment by fast rotaƟonal (liŐhthouse like) movements͘ 3D imaŐinŐ in this case is achieved by a pullback of the probe͘

(30)

Chapter 2 28

RESULTS

Skin Cancer

Non-melanoma skin cancer

Non-melanoma skin cancer (NSDC), includinŐ basal cell carcinomas (BCC), sƋuamous cell carcinomas (SCC) and the premaliŐnant acƟnic keratosis (<), is the most prevalent cancer in liŐht-skinned populaƟons ΀12΁͘ OCT could beneĮt the diaŐnosƟc manaŐement of NDSC, especially in paƟents with mulƟple lesions or lesions at locaƟons where resecƟon results in bad cosmeƟc outcome ΀13΁͘

Several studies are performed to assess whether OCT is useful in diaŐnosinŐ tumors and deĮninŐ tumor borders ΀14-20΁͘ Olmedo et al͘ invesƟŐated 27 paƟents with BCC͘ In 20 of them the OCT imaŐes matched with histopatholoŐy͘ typical OCT imaŐe of BCC is shown in FiŐ͘ 3a͘ In these imaŐes, the dark lobules represenƟnŐ BCC appeared as a reŇecƟon density of the epidermis͘ decreased reŇectance was seen around the periphery of the lobules of the tumor, while the Įbrous stroma closely surroundinŐ the BCC showed increased intensity͘ For the seven specimens that did not match, technical issues had interfered with the eǆaminaƟon ΀18΁͘ 'ambichler et al͘ imaŐed BCC as well as healthy skin͘ In OCT imaŐes, healthy skin showed layered skin architecture, whereas BCC showed loss of these layers and disarranŐement of the epidermis and dermis ΀14΁͘ bsence of well-deĮned layerinŐ in NDSC has been seen by others as well ΀16, 20΁͘ DoŐensen et al͘ compared OCT directly to the accuracy of patholoŐy diaŐnosis͘ DependinŐ on the observers, sensiƟvity and speciĮcity varied from 57 to 94й and 43 to 96й, respecƟvely͘ Eǆperienced observers reached a sensiƟvity of 79 to 94й and a speciĮcity of 85 to 96й ΀16΁͘ Instead of observinŐ the OCT-imaŐes, :orŐensen et al͘ implemented a machine-learninŐ analysis͘ Correct classiĮcaƟon of 37 < lesions and 41 BCCs compared to patholoŐy were made with an accuracy of 73й (<) and 81й (BCC) ΀15΁͘ Other studies focused on measurinŐ the thickness of BCC ΀19΁ and < ΀17΁ in vivo͘ stronŐ correlaƟon between tumor thickness measured in both OCT and patholoŐy was found (pф0͘001) ΀19΁͘ Overall, OCT is able to show the diīerent skin layers and is capable to measure tumor thickness in NDSC͘

Malignant Melanoma

In case of a maliŐnant melanoma (DD), thickness is the most important factor for (a) staŐinŐ, (b) determininŐ the siǌe of the eǆcision marŐins, and (c) the need for senƟnel lymph node biopsy͘ hsinŐ OCT to measure DD thickness could enable beƩer pre-operaƟve planninŐ by deĮninŐ resecƟon marŐins and the need for senƟnel lymph node biopsy͘ Doreover, a non-invasive diaŐnosƟc tool capable of disƟnŐuishinŐ between beniŐn nevi (BN) and DD would decrease the amount of unnecessary harvested biopsies and eǆcisions͘ HistopatholoŐically, DD consists of atypical melanocytes (with atypical nuclei and nucleoli) that invade the epidermis and someƟmes even invade the dermis͘ These atypical cells may have a diīerent nuclear refracƟve indeǆ and thus lead to diīerence in scaƩerinŐ ΀11΁͘ typical OCT imaŐe of DD is shown in FiŐ͘ 3b͘

(31)

2

(32)

Chapter 2 30

Only a few studies are performed to invesƟŐate whether OCT is useful in detecƟnŐ DD͘ De 'iorŐi et al͘ studied the piŐment network and brown Őlobules in melanocyƟc piŐmented lesions as seen with dermatoscopy ΀21΁ and compared these features with OCT and histoloŐy͘ In 6 out of 10 cases it was possible to compare structures found in dermatoscopy and histoloŐy to structures in OCT͘ However, it was impossible to diīerenƟate DD from BN͘ On contrary, a study by 'ambichler et al͘ showed siŐniĮcant diīerences in OCT imaŐinŐ of BN and DD ΀22΁͘ total of 75 paƟents with 52 BN and 40 DD were included͘ DD showed a marked architectural disarray (pс0͘036) and rarely displayed a clear dermo-epidermal border (pс0͘0031) when compared to BN͘ The most prominent OCT feature of DD was larŐe, verƟcal cone-shaped structures, which were not observed in BN (Pф0͘001) ΀22΁͘ lthouŐh these results are promisinŐ, a larŐe clinical study is needed to establish robust diīerenƟaƟon of melanoma from beniŐn nevi͘

Oral͕ laryngeal͕ anĚ esophageal cancer

Oral cancer

SƋuamous cell carcinoma (SCC), accounƟnŐ for 96й of all oral cancers, is mostly preceded by dysplasia presenƟnŐ as white or red lesions on the oral mucosa (leukoplakia, erythroplakia)͘ Currently, these potenƟally maliŐnant lesions are biopsied or eǆcised and therefore less suitable for screeninŐ of hiŐh-risk populaƟon͘ Only a few studies have been performed on OCT in the oral cavity͘ One of these studies included 143 healthy volunteers and demonstrated a varyinŐ thickness of epithelium dependinŐ on its locaƟon within the oral cavity͘ The larŐest thickness values were measured in the reŐion of the buccal mucosa and the hard palate, whereas the thinnest epithelium was found at the Ňoor of the mouth ΀23΁͘ Wilders-Smith et al͘ imaŐed 50 paƟents with oral suspicious lesions with OCT, shown in FiŐ͘ 3c͘ Ōer imaŐinŐ, standard biopsy and histopatholoŐy were performed͘ Two invesƟŐators diaŐnosed the lesions blinded on OCT and histopatholoŐy subseƋuently͘ For detecƟnŐ carcinoma in situ or SCC versus non-cancer, sensiƟvity was 0͘931 and speciĮcity was 0͘931; for detecƟnŐ SCC versus all other patholoŐies, sensiƟvity was 0͘931 and speciĮcity was 0͘973 ΀24΁͘ nother study showed diīerences between normal oral mucosa, dysplasia and SCC scorinŐ the OCT-imaŐes on three indicators: epidermal layer thickness, standard deviaƟon of an -mode scan intensity proĮle in the epidermal layer and the aƩenuaƟon of liŐht͘ However, the clinical value of this study is limited due to unclear methodoloŐy ΀25΁͘ Recently, deŐun et al͘ measured the backscaƩered liŐht intensity as a funcƟon of depth in OCT-imaŐes of biopsy samples from the oral cavity͘ Hereby, they could diīerenƟate dysplasia from normal control samples ΀26΁͘ Overall, OCT seems to diīerenƟate well between normal Ɵssue and SCC͘

Laryngeal cancer

Several studies in the human larynǆ have been performed͘ In a study of 82 paƟents underŐoinŐ surŐery for various head and neck patholoŐy, OCT provided informaƟon on epithelial layer thickness, inteŐrity of the basement membrane (BD), and microstructural features͘ Whereas

(33)

2

the normal epithelial thickness varied from 98 to 185 ђm, the epithelial layer was thickened up to 300 ђm in hyperkeratoƟc lesions͘ In hyperkeratoƟc lesions, the BD on OCT imaŐes was sƟll intact in contrary to early microinvasive SCC of the larynǆ, where disrupƟon of the BD was seen on OCT imaŐes͘ Doreover, OCT-imaŐes were comparable with convenƟonal histopatholoŐy ΀27΁͘

Esophageal cancer

Over the past two decades, the incidence of esophaŐeal adenocarcinoma (C) has risen considerably, especially in the distal esophaŐus and the esophaŐoŐastric juncƟon͘ Several studies showed that Őastro-esophaŐeal reŇuǆ disease ('ERD) increases the risk of ŐeƫnŐ esophaŐeal C ΀28΁ by promoƟnŐ the formaƟon of Barret͛s esophaŐus (BE), a precursor of C͘ t the moment, endoscopy and random taken biopsies are used to follow up paƟents with BE͘ Endoscopic OCT miŐht be helpful in selecƟnŐ suspicious areas in BE for further biopsy͘ The use of OCT in imaŐinŐ the esophaŐus and the esophaŐoŐastric juncƟon has been widely studied͘ Eǆ-vivo studies showed that OCT can diīerenƟate diverse layers in the esophaŐeal wall ΀29΁ and can Őenerate imaŐes correspondinŐ to histoloŐy ΀30-31΁͘ Poneros et al͘ formulated diīerent criteria for intesƟnal metaplasia on the basis of OCT imaŐes͘ When tested prospecƟvely, these criteria had a sensiƟvity of 97й and a speciĮcity of 92й ΀32΁͘ >ikewise, others acƋuired OCT-imaŐes of the esophaŐus and stomach in 69 paƟents containinŐ normal sƋuamous mucosa, BE and C͘ These OCT-imaŐes were accurately recoŐniǌed by observers unaware of their site of oriŐin, with an accuracy of 84͘6й ΀33΁͘ IsenberŐ determined the accuracy of endoscopic OCT in diaŐnosis of dysplasia in paƟents with BE͘ They reported an accuracy of 78й for the detecƟon of dysplasia in BE͘ Dain limitaƟons of the study were the variability in endoscopists͛ accuracy rates, diĸculty in real-Ɵme interpretaƟon and the need for reĮned criteria of dysplasia in endoscopic OCT imaŐinŐ ΀34΁͘ OCT of dysplasia was studied in more detail by Evens et al͘ They studied biopsy-correlated OCT-imaŐes to establish OCT imaŐe characterisƟcs able to diīerenƟate intramucosal carcinoma (ICD) and hiŐh-Őrade dysplasia (H'D) from low-Őrade dysplasia (>'D), because ICD and H'D have a hiŐher risk of developinŐ into C͘ There was a siŐniĮcant relaƟonship between the histopatholoŐical diaŐnosis of IDCͬ H'D and the scores for each OCT imaŐe feature ΀35΁͘ Recent improvements in endoscopic OCT made it possible to imaŐe larŐer areas such as the enƟre distal esophaŐus (approǆimately 6͘0 cm) with ultrahiŐh-resoluƟon ΀36-37΁͘ n eǆample of a typical OCT imaŐe of the esophaŐus is Őiven in FiŐ͘ 3d͘ With this approach, it may be possible to accurately deĮne paƟents with dysplasia who are at risk for developinŐ C͘

Genital anĚ ďlaĚĚer cancer

sƵlǀar cancer

In the last 30 years the incidence of VIN (vulvar intraepithelial neoplasia), a premaliŐnant skin disorder that oŌen causes pruritus, pain and psychoseǆual dysfuncƟon, has increased more than 400й to approǆimately 2͘5 cases per 100,000 women in the hS͘ The proŐression rate

(34)

Chapter 2 32

of VIN into VSCC is about 9й in untreated paƟents and 3͘3й in paƟents aŌer treatment͘ VIN is treated usinŐ conservaƟve surŐical eǆcision, laser vaporiǌaƟon, or medical therapy͘ PaƟents are reŐularly eǆamined to foresee occult invasion and check for possible new VIN lesions ΀38΁͘ DeĮnite diaŐnosis of a vulvar lesion of uncertain siŐniĮcance is obtained by punch biopsy͘ However, every aƩempt is made to avoid vulvar muƟlaƟon that may lead to psychoseǆual distress ΀39-40΁͘ Hence, in the present workŇow in the diaŐnosis and treatment of VIN, there is a clear need for non-invasive diaŐnosƟc tools͘

In a study by Escobar et al͘, three paƟents with PaŐet͛s disease (a potenƟal premaliŐnant lesion) were imaŐed with OCT ΀41΁͘ The authors observed clear irreŐulariƟes in the epithelial layer, and disrupƟon of the basal membrane͘ Besides disrupƟon of the basement membrane, OCT may be helpful in characteriǌaƟon of vulvar lesions by demonstraƟnŐ diīerences in layer thickness and in the aƩenuaƟon of liŐht between diīerent Ɵssues͘ Data from our Őroup in 16 paƟents with VIN show a siŐniĮcant diīerence in the thickness of the epidermal layer as well as in the aƩenuaƟon of liŐht between VIN lesions and healthy skin ΀10΁͘ These diīerences probably occur because of the Őrowth of neoplasƟc cells and chanŐed nuclearͬcytoplasma raƟo in VIN-lesions compared with healthy cells ΀11΁͘ These data indicate the potenƟal of OCT to disƟnŐuish between healthy and VIN lesions͘ In FiŐ͘ 3e, an eǆample of an OCT imaŐe of a VIN lesion is shown͘

erǀical cancer

Cervical cancer is Őenerally preceded by intraepithelial neoplasia (CIN)͘ Hence, the implementaƟon of new imaŐinŐ techniƋues that allow cheap, real-Ɵme, and non-invasive detecƟon of premaliŐnant abnormaliƟes have the potenƟal to improve the loŐisƟcs and economics of screeninŐ proŐrams for invasive cervical cancer ΀41-42΁͘

The diaŐnosƟc value of OCT in the detecƟon of CIN has been studied by several Őroups͘ In 1999, the Įrst eǆ vivo study on 32 cervical specimens concluded that an intact basement membrane - a feature of healthy epithelium in histoloŐy - can be seen with OCT ΀43΁͘ The Įrst in vivo studies were performed by Escobar et al͘ ΀41΁͘ OCT-imaŐes of paƟents with CIN I, II, and III lesions were directly compared to histoloŐy reports͘ ll imaŐes of normal cerviǆ showed a clear epithelial layer and a basal membrane that was deĮned as an interface between a briŐhter scaƩerinŐ (stroma) and a poorer scaƩerinŐ (epithelium) reŐion in the OCT imaŐe͘ The authors concluded that this paƩern presented normal sƋuamous epithelium͘ In 16 of the 18 paƟents with CIN II andͬor III, the imaŐes showed an unstructured homoŐenous area with hiŐhly backscaƩerinŐ reŐion and fast aƩenuaƟon of the siŐnal, however this was not ƋuanƟĮed͘ This correlaƟon between the intensity of the backscaƩerinŐ liŐht from the epithelia of normal and abnormal Ɵssue in the cerviǆ, was also found by others ΀44΁͘ To determine further accuracy of the OCT imaŐes, a ŐradinŐ scale was used to describe the OCT-imaŐes: normal if a well-orŐaniǌed 2-layered structure was seen with a sharp interface between the surface epithelium (sƋuamous) and underlyinŐ layer (connecƟve Ɵssue); abnormal if the Ɵssue was unstructured

(35)

2

with no interface present and intermediate if irreŐulariƟes on the imaŐes suŐŐested arƟfacts or physioloŐical condiƟons that did not meet the criteria͘ With this ŐradinŐ scale, sensiƟvity and speciĮcity of OCT alone for ш CIN II lesions was 56 and 59й, respecƟvely ΀45΁͘

If paƟents with posiƟve cerviǆ cytoloŐy for dysplasƟc cells underŐo colposcopy, OCT miŐht have the potenƟal to predict if a speciĮc spot seen with colposcopy is CIN I, II, III or invasive cancer͘ In this seƫnŐ, >iu invesƟŐated OCT imaŐinŐ in colposcopy͘ OCT decreased the sensiƟvity to detect ш CIN II lesions form 63 to 36й, however speciĮcity increased from 83 to 93й ΀46΁͘ To improve clinical analysis, >ee looked at the deŐrees of circular polariǌaƟon caused by the scaƩerinŐ chanŐes induced by CIN͘ linear ĮƫnŐ procedure was used to ƋuanƟfy the aƩenuaƟon of the deŐrees-of-polariǌaƟon siŐnal͘ This approach resulted in sensiƟvity and speciĮcity to detect CIN of 94͘7 and 71͘2й, respecƟvely ΀47΁͘ n OCT imaŐe of a CIN I lesion is shown in FiŐ͘ 3f͘ Further improvements in OCT resoluƟon are currently tested to improve accuracy of OCT for diaŐnosinŐ cervical cancer and CIN͘

Bladder cancer

In non-diseased Ɵssue, the three anatomic layers of the bladder wall (urothelium, lamina propria and muscularis propria) can be well disƟnŐuished with OCT ΀48-49΁͘ Hermes et al͘ evaluated OCT in 142 human bladder specimens eǆ vivo and demonstrated that OCT could discriminate between normal, CIS and invasive transiƟonal cell carcinoma (TCC) with a sensiƟvity of 83͘8й and a speciĮcity of 78͘1й͘ None of the CIS or TCC samples was classiĮed as normal Ɵssue based on the OCT imaŐe (no false-neŐaƟves), 6 samples of normal Ɵssue were classiĮed as TCC (false-posiƟve rate 5͘7й) and 24 samples could not be classiĮed (16͘9й) ΀50΁͘ Two Őroups invesƟŐated the diaŐnosƟc accuracy of cystoscopic applied OCT͘ Danyak et al͘ classiĮed bladder Ɵssue samples as beniŐn or maliŐnant with an overall sensiƟvity and speciĮcity of 100 and 89й, respecƟvely͘ Ten maliŐnancies were classiĮed by OCT as invasive (шT1), of which nine were conĮrmed by histoloŐy ΀51΁͘ >erner et al͘ showed that diīerenƟaƟon of bladder tumours conĮned to the mucosa (Ta) was possible with a sensiƟvity of 90й and a speciĮcity of 89й͘ For diīerenƟaƟon of T1 and T2 tumours, sensiƟvity was 75 and 100й, while speciĮcity was 97 and 90й ΀52΁͘ False-posiƟve ĮndinŐs could Őenerally be eǆplained by inŇammatory lesions of the bladder ΀51΁͘ >inŐley-Papadopoulos et al͘ developed an automated alŐorithm that allows recoŐniƟon of teǆture within an OCT imaŐe to provide the uroloŐist with a diaŐnosis͘ This alŐorithm was able to diīerenƟate beniŐn and maliŐnant bladder Ɵssue with a sensiƟvity of 92й and a speciĮcity of 62й ΀53΁͘

Since real-Ɵme hiŐh resoluƟon OCT imaŐes can be obtained durinŐ cystoscopy, the techniƋue may especially be useful for Őuidance of biopsy procedures and staŐinŐ of suspected Ɵssue areas within the bladder͘ In FiŐ͘ 3Ő, an OCT-imaŐe of a papillary carcinoma of the bladder is shown͘

(36)

Chapter 2 34

DISCUSSION AND CONCLUSIONS

OCT is an opƟcal diaŐnosƟc tool that aims to predict in vivo histopatholoŐic diaŐnosis in a non-invasive way͘ It produces real-Ɵme hiŐh-resoluƟon imaŐes comparable with histopatholoŐy͘ Several studies have shown that OCT can be used to disƟnŐuish healthy skin from NDSC͘ For detecƟon of melanoma lesions, however, the accuracy is sƟll limited͘ In the oral cavity and larynǆ, OCT is able to disƟnŐuish between normal Ɵssue and SCC͘ In addiƟon, OCT can Őive informaƟon about Ɵssue thickness and basement membrane inteŐrity͘ DiīerenƟaƟon between dysplasia and normal can sƟll be improved͘ In the esophaŐus, larŐe areas with hiŐh resoluƟon can be imaŐed allowinŐ Ɵssue characteriǌaƟon of dysplasia as well as tumor Ɵssue͘ For vulva lesions, OCT proved to be able to diīerenƟate between healthy Ɵssue and VIN lesions, while in the cerviǆ OCT allowed diīerenƟaƟon between normal cervical Ɵssue and CIN͘ However, OCT is not capable of disƟnŐuishinŐ diīerent CIN Őrades͘ Bladder OCT so far has shown the potenƟal to diīerenƟate Őrade and staŐe in small populaƟons but it needs a larŐer populaƟon study to provide deĮnite answers͘

FƵtƵre perspecƟǀes

OCT holds the potenƟal to provide ĨƵncƟonal opƟcal biopsies of epithelial cancers by combininŐ imaŐinŐ with ƋuanƟĮcaƟon of physioloŐical funcƟonal parameters, e͘Ő͘, perfusion and oǆyŐenaƟon or cellular orŐaniǌaƟon͘ This can be achieved either by advanced analysis of the OCT siŐnal itself ΀54-55΁ or by combininŐ OCT with other imaŐinŐ modaliƟes such as Raman spectroscopy (RS) ΀56΁ or Ňuorescence spectroscopy ΀57΁͘ nalysis of the spaƟal and temporal chanŐes of the OCT siŐnal allows determinaƟon of the opƟcal aƩenuaƟon coeĸcient (which is related to Ɵssue orŐaniǌaƟon) and blood perfusion͘ We have demonstrated that the aƩenuaƟon coeĸcient discriminates between normal and (pre)cancerous lesions in vulva ΀10΁, bladder ΀9΁, and kidney ΀7-8΁͘ dvanced implementaƟons of OCT use liŐht that is eǆtended from the near infrared to the visible wavelenŐths, allowinŐ ƋuanƟĮcaƟon of hemoŐlobin concentraƟons ΀58΁ and oǆyŐen saturaƟon ulƟmately allowinŐ to map oǆyŐen metabolism of individual locaƟons in the lesion͘ Complementary informaƟon on biochemical composiƟon can be obtained by combininŐ OCT with RS͘ In that way, RS can esƟmate the molecular composiƟon of Ɵssue, while OCT can produce imaŐes͘ The use of OCT in combinaƟon with other techniƋues can hereby assist in screeninŐ and eventually even diaŐnose maliŐnancies non-invasively͘

AcknoǁleĚgements

The authors would like to acknowledŐe Swanson ΀18΁, Wilder-Smith ΀24΁, Suter ΀36΁, and 'allwas ΀42΁ and their publishers for providinŐ them OCT ĮŐures of their studies͘