Fluorescence Excitation Emission Matrices of Human Tissue: A ...

F lu orescen ce E x cita tio n E m issio n M a trices o f H u m a n T issu e: A S y stem fo r in Vivo M ea su rem en t a n d M eth o d o f D a ta A n a ly sis A N D R EÂS F . Z U L U A G A , U R S U T Z IN G E R , A N T H O N Y D U R K IN , H O L G E R F U C H S , A N N G IL L E N W A T E R , R H O N D A JA C O B , B O N N IE K E M P , J A M E S F A N , a n d R E B E C C A R IC H A R D S -K O R T U M * D ep art m en t o f E lectrical and C om puter E n gineerin g, T he U niversity o f T exas at A ustin, A ustin, T exas 78712 ( A .F .Z ., U .U ., H .F ., R .R .- K.) ; Ca nd ela Co rp ora tio n, W aylan d, M assa chu se tts (A .D .); and D epartm ent of H ead and N eck Surg ery , T h e U T M D A nd erso n Ca n cer C enter, H o usto n, T exa s 7 7 030 ( A .G ., R .J ., B .K .)

W e d esc rib e a sy ste m ca p a b le of m easu rin g sp at ially reso lved re¯ ecta n ce sp ectra fro m 3 8 0 t o 9 50 n m an d ¯ u oresc en ce ex citat ion em issi on m a t rices f ro m 3 30 to 5 00 n m excitatio n an d 380 to 7 00 n m em issi on in vivo . S yste m p erf orm a n ce w as com p a red t o t h a t of a st an d ard sc a n n ing sp ectro ¯ u orim eter. T h is `` F ast EE M `` sy st em w a s u sed to int errog a te h u m an n orm a l an d n eop lasti c oral cavity m u cosa in vivo . M easu rem en ts w ere m a d e th rou gh a ® b er- op t ic p rob e an d req u ire 4 m in to t a l m easu rem en t tim e. W e p rese n t a m et h od b ase d on au t oco rrela tio n v ectors t o id en t ify ex cita t ion an d em issio n w avelen gt h s w h ere t h e sp ectra of n o rm al an d p at h o lo g ic t issu es d iffer m ost. T h e F a st EE M sy ste m p ro v ides a t o ol w ith w h ich to st u d y th e relat ive d iag n o sti c a b ility o f ch an g es in a b so rp tion , sc at terin g, an d ¯ u orescen ce p rop ert ies of tissu e. In dex H eading s: F lu oresc en ce; R e¯ ect a n ce; EE M ; T issu e; H ea d an d n eck ; N eo p lasia .

IN T R O D U C T IO N A g ro w in g n u m b er o f clin ical stu d ies h ave d em o n strated th at ¯ u o rescen ce spectro sco p y can b e used to distingu ish n o rm al an d ab n o rm al h u m an tissu es in vivo in th e skin , 1 head an d neck , 2,3 gen ito -urin ar y tract,4 ,5 g astro -in testin al tract, 6 ,7 b reast, 8 an d b rain . 9 R eferen ces 1 0 ±1 3 pro vid e recent review s o f th is ® eld . It is w ell k n ow n th at ¯ u o rescen ce inten sity an d line sh ap e are a fun ctio n o f bo th th e excitatio n and em issio n w avelen gth in sam p les con tain in g m u ltip le ch rom o ph o res, such as h u m an tissue. A co m p lete characterizatio n o f th e ¯ u o rescen ce p ro p erties of an un k n ow n sam ple req uires m easu rem ent o f a ¯ u o rescen ce excitatio n em issio n m atrix , in w hich th e ¯ u o rescen ce in ten sity is reco rded as a fu n ctio n o f b oth excitatio n an d em issio n w av elen g th. T h e ® eld of an aly tical chem istr y has exp lo ited th e ¯ uo rescen ce p ro p erties of d ifferen t com po u n ds to id en tify an d qu an tify th em in m ix tu res. H o llan d et al. 1 4 described a com p uter-b ased system to m easure ¯ uo rescen ce and ab sorb ance sim ultaneo u sly fro m qu in in e b isulfate. W arner an d co lleag ues 15 bu ilt an in strum ent to collect a m atrix o f 24 1 ¯ u o rescen ce spectra at 24 1 d ifferent excitatio n w aveleng th s in 1 6 .7 m s an d used it to analy ze the ¯ u o rescen ce pro perties o f m ix tu res of h y dro carb o ns. L in ear algeb ra m eth od s w ere th en ap plied to d eterm in e qu an titativ e con centratio n and nu m ber of analy te d ata fro m these m easurem en ts. 1 6±1 9 T h ese m etho d s assum e a lin ear sy stem and are m u ch m o re su ccessfu l in n o n scatterin g , o ptically d ilu te sam ples R eceiv ed 16 Ju ly 1 99 8; accep ted 3 N ovem ber 19 98. * A uthor to w h om co rr esp o nd en ce sh o uld be sen t.

302

Volume 53, Number 3, 1999

th an th ey are in h u m an tissu e. In tissu e, ch rom o ph o res are em b edd ed in a scatterin g an d ab sorb in g m ed iu m affecting b o th th e ex citatio n ligh t reach ing th e ch rom o ph o res an d th e em issio n ligh t leavin g the sam p le. T h e relatio n ship b etw een th e ¯ u o rescen ce spectra o f d ilu te and turb id sam ples h as been exp lored 20 and fou n d to b e hig h ly n o nlinear. S tud ies h av e also been carried ou t in vestigatin g th e p erform ance of different lin ear alg ebra m etho d s in the d eterm ination o f co ncen tration s an d n u m ber o f chro m o p ho res presen t in a ho m o g en eo u s tu rbid sam p le. 2 1 H ow ev er, it is n ot clear th at they can b e ap plied to inh o m o g eneo u s tissu es. M ost clin ical stu dies rep o rted to date h ave m easu red ¯ u o rescen ce em issio n sp ectra at o nly a sm all n u m b er o f excitatio n w avelen gth s (ty p ically on e to three) d u e to clin ical req uirem en ts im p osed o n th e size, speed , an d sen sitiv ity o f in stru m en tatio n. T h e ch oice o f ex citatio n w av elen g th h as b een b ased o n facto rs w hich v ar y from stu d y to stu d y, b u t in clu d e laser av ailab ility, 2 2 pred ictio ns of ch ro m o p h ores th o u gh t to b e presen t in n o rm al an d abn o rm al tissu es, an d m easurem en ts o f ¯ u o rescen ce excitatio n em ission m atrices (E E M s) o f no rm al and abn o rm al tissues in vitro. 2 3,24 W hile in vitro m easurem en ts o f tissu e E E M s are feasible w ith th e use o f co m m ercially available scan n in g ¯ uo rim eters, sev eral stu dies h av e dem on strated th at th e op tical p ro perties o f tissu e ch ang e sig n i® can tly w hen tissue is ex am in ed in vitro du e in p art to in terru p tio n of the b lo o d su p ply, 2 5 o x id ation , 2 6 an d sm all size o f b iop sies. 2 7 T hu s, in vitro stu d ies to select excitatio n w av elen g th s are o f lim ited v alu e. It is w ell k no w n th at the ab sorp tio n and scatterin g pro perties o f tissu es in vivo affect bo th the inten sity an d lin e sh ap e o f m easu red ¯ u orescence sp ectra. 28 T h e o ptical pro perties o f tissu e var y sp atially, d epen d in g o n its architectu re, b loo d sup p ly, and m etab o lic state; h o w ev er, spatially reso lv ed m easurem en ts o f d iffusely re¯ ected lig h t can b e u sed to estim ate tissu e op tical p rop erties in vivo . 2 9 ±3 3 S everal recen t stud ies h ave su g g ested that differences in th ese o p tical p ro p erties, assessed b y u sin g d iffu se re¯ ectance sp ectro scop y, can b e used to d iscrim in ate no rm al and ab n orm al h u m an tissu es in vivo in th e u rin ar y blad der 34 an d th e sk in . 3 5 F u rth erm o re, m easuring b oth ¯ u o rescen ce and diffu se re¯ ectance sp ectra m ay p ro vid e add itio n al in fo rm atio n o f d iagn o stic v alu e. 36 ,37 A system cap able of m easu rin g sp atially reso lv ed re¯ ectance sp ectra and ¯ uo rescen ce ex citation em issio n m atrices in vivo w o uld rem o v e the lim itation s o f m an y

0003-702 8 / 99 / 5303 -0 302$2.0 0 / 0 q 1999 S ociety for A pplied S pectros copy

APPLIED SPECTROSCOPY

p rev iou s stu d ies, p oten tially en abling pred ictio n o f tho se ex citatio n w av elen g th s w h ich p ro v ide g reatest d iscrim in atio n o f n o rm al an d ab no rm al tissu es, as w ell as a b etter u nd erstan d ing o f th e relativ e d iagn o stic ability o f chan g es in abso rp tio n, scattering , an d ¯ uo rescen ce p rop erties o f tissue. Tw o ® b er-o ptic system s to reco rd ¯ uo rescen ce E E M s an d re¯ ectan ce sp ectra at a sin g le sp atial location h ave been prev io u sly described in the literatu re. 38 ,39 Z u clich an d co lleag ues 3 9 d evelo ped a ® b er-o ptic sy stem to reco rd b o th d iffusely re¯ ected ex citation ligh t and ¯ uo rescen ce em issio n b y usin g a sin gle 1 4 b it m u ltichan n el d etecto r. S equ en tial m easu rem en ts w ere m ad e, ® rst w ith w hite lig h t to collect th e d iffuse re¯ ectan ce spectrum an d then at a series of excitatio n w av elen g ths to co llect b o th elastic b ackscatterin g an d any ind u ced ¯ uo rescen ce. L ig h t w as pro vid ed by a p ulsed , ® ltered arc lam p fro m 3 00 to 6 00 n m in 1 0 nm in crem en ts, an d fo llo w in g m easu rem ent, all sp ectra w ere assem b led to y ield th e diffu se re¯ ectan ce spectru m and a ¯ uo rescen ce E E M for a sin g le sp atial lo catio n . C o llectio n o f elastic back scatterin g an d ¯ uo rescen ce E E M s fro m no n scatterin g tissu es such as th e h um an len s requ ired 1±2 m in. Ty p ically, the elastic b ack scattering w as at least 3± 4 o rders o f m agn itud e stro n ger than th e d etected ¯ u orescence, leav in g a v er y lim ited d y n am ic rang e av ailab le to m easu re tissue ¯ uo rescence. M o re recen tly, Z an g aro et al. 38 d escrib ed a sy stem to reco rd ex citation em issio n m atrices an d diffu se re¯ ectance fro m a sin gle sp atial lo catio n . A N 2 laser that p um p s a series o f 1 0 d y e-co n tain in g cu vettes m ou n ted o n a rotating w h eel p ro v id es seq uen tial ex citatio n at 1 1 d ifferen t w aveleng th s fro m 3 37 to 50 0 nm ; m u ltich an nel d etection at each ex citatio n w av elen gth yields 1 1 em issio n sp ectra that are th en assem bled to y ield th e ex citation em ission m atrix . W h ite ligh t illu m in atio n is th en u sed to p rod u ce a sin g le d iffu se re¯ ectan ce sp ectru m . C o llectio n o f a tissu e E E M and re¯ ectan ce sp ectru m req uires ap p ro x im ately 60 0 m s at a sig nal-to -no ise ratio (S N R ) of 5 0:1 fo r colo n tissu e. In th is p ap er w e d escrib e a system fo r in vivo m easu rem ent of ¯ u orescence ex citatio n em issio n m atrices an d sp atially reso lved d iffu se re¯ ectance spectra. A n arc lam p co u pled to a scan n ing sp ectro m eter p ro v ides co n tinu o usly tu nab le ex citatio n lig ht, w h ich is co u pled in to a ® ber-o p tic p ro b e. R esulting tissu e ¯ u orescence is collected thro ug h th e ® b er p ro b e an d d eliv ered to an im ag ing sp ectro grap h an d ch arg e-co u pled d ev ice (C C D ) cam era. F lu o rescen ce em ission sp ectra are co llected sequ en tially at 1 8 excitatio n w av elen g ths ran g ing from 3 3 0 to 5 00 nm , w h ich are th en assem b led in to a ¯ uo rescen ce E E M . S u bseq u ently, w hite ligh t is cou p led in to th e p rob e, an d diffu sely re¯ ected ligh t ex itin g th e tissu e at th ree sp atial lo catio n s is detected fro m 38 0 to 9 50 n m by u sin g the spectro g rap h and C C D . W ith th is system , tissue E E M s an d re¯ ectan ce sp ectra w ith S N R in ex cess o f 25 : 1 can b e collected in ap p ro x im ately 4 m in . W e p resen t E E M s an d re¯ ectan ce sp ectra o f n o rm al hu m an o ral cavity m u cosa an d a tu m o r o f th e o ral cav ity ob tain ed in vivo. F in ally, w e describe a new m eth od , b ased o n ex citatio n and em issio n au to co rrelatio n v ectors, to analy ze ¯ uo rescen ce E E M s to d eterm ine excitatio n and em issio n w avelen gth regio n s w h ere differen ces b etw een no rm al an d abn o rm al tissu es are g reatest. W e illu strate th e ap p lication o f th is m eth od to th e E E M s.

F IG . 1. S ystem b lock d iag r am sh ow ing a var iable excitation ligh t so urce, a ® ber-o ptic deliv er y and collection pr obe, and a sp ectral m u ltichann el analyzer.

M A T E R IA L S A N D M E T H O D S F igu re 1 illu strates a b lo ck d iagram o f the sy stem (F astE E M sy stem ). T h e sy stem co nsists of th ree m ain co m p on en ts: (1 ) an arc lam p , step per m o tor-d riven m o n och ro m ato r, and ® lter w h eel, w h ich p ro v ides m o no ch ro m atic an d bro ad ban d ex citatio n ; (2 ) a ® b er-o ptic p rob e that d irects ex citatio n lig ht to th e tissu e an d collects rem itted ¯ u orescence fro m on e lo catio n an d d iffu sely re¯ ected ligh t fro m three lo catio ns; an d (3) a ® lter w h eel, im ag ing sp ectro grap h, an d C C D cam era th at d etects th e sp ectrally reso lv ed re¯ ectan ce an d ¯ u orescence sig nals. E x citatio n m o n och rom ator p osition , ® lter w h eel p o sitio n , sp ectro grap h g ratin g p o sitio n, C C D op eratio n, an d data acqu isitio n are co n tro lled by a lap top p erson al co m p u ter m ated to a d ock ing statio n . T h e sp eci® catio ns of each su b-sy stem are describ ed b elo w . T h e p rob e, illustrated in F ig . 2 , co n sists of a total o f 4 6 o p tical ® b ers [2 00 m m diam eter, n um erical ap ertu re (N A ) 5 0 .2 ] arran g ed in tw o co n centric bu n dles. T he center bu n d le co n tain s 25 ¯ uo rescen ce ex citation ® bers an d 12 ¯ u orescen ce co llection ® b ers. T he pro xim al en d s o f th e ¯ u orescence excitatio n ® b ers are arrang ed in tw o v ertical lin es at th e ex it slit o f th e ex citation m o no ch ro m ato r to m ax im ize th e co up ling of th e ligh t in to th e sam p le. T h e pro x im al en d s o f th e ¯ u o rescen ce co llection ® b ers are arrang ed in a sing le v ertical line at th e en tran ce slit of the im ag in g sp ectro g rap h . A t th e d istal en d o f th e p ro b e, th e ® b ers that ex cite and co llect ¯ uo rescen ce are arran ged rand o m ly in a cen tral bu n dle an d placed in co n tact w ith a sho rt piece o f a th ick q u artz ® b er (2 m m d iam eter, 1 5 m m lo ng , N A 5 0 .2 ). T h e d istal tip o f this ® ber is p laced in co ntact w ith the sam ple su rface an d en sures th at th e area from w hich ¯ u o rescen ce is co llected is th e sam e as th at directly illu m in ated . T h e n ine ® bers fo r illum ination an d co llection o f diffuse re¯ ectan ce are arran g ed in a co ncen tric ring aro un d the thick q u artz ¯ u o rescen ce m easurem en t ® b er. T h e d istal end s o f these ® b ers are ¯ u sh w ith th e tip o f th e cen tral ® ber an d are p laced in co ntact w ith th e sam p le su rface.


303

F IG . 2. (L eft) S ch em atic d iag ram of th e d istal end s of the pro be: (a) ou ter sh aft, (b ) ¯ u or escence ex citation and em issio n ® bers, ( c ) r e¯ ectance collection an d illum inatio n ® b ers, (d ) m ixing elem en t, (E ) r e¯ ectance excitation ® ber, (1 ±3 ) re¯ ectan ce collection location s; ( rig ht) sc hem atic diagram o f th e pro x im al ends of the p ro be.

W h ite ligh t fro m a p o rt on the side o f th e lam p h o usin g is co up led to th e pro xim al en d o f a sin g le illu m in atio n ® b er (8 0 m m , N A 0 .2 ). P ho to ns th at scatter th ro u gh th e tissu e and ex it the su rface are co llected at ® ve d ifferen t po sitio n s w ith sev en collectio n ® bersÐ th ree lo cated 1 8 0 8 fro m th e illu m in atio n ® b er (3 m m distance), tw o lo cated 90 8 from the illu m in atio n ® b er (2 .1 m m ), and tw o lo cated 45 8 fro m the illu m inatio n ® b er (1 .1 m m ) as sh ow n . T h e pro xim al end s o f th e re¯ ectan ce co llectio n ® b ers are situated at th e top of th e v ertical lin e o f ¯ u o rescen ce collection ® b ers, sep arated b y du m m y ® bers, as sho w n in F ig . 2. T h e lig h t sou rce for the instrum en t, w hich pro vid es bo th q u asi-m o n och rom atic excitatio n fo r ¯ u o rescen ce and b ro ad ban d illu m inatio n for re¯ ectan ce, is a 1 5 0 W ozo n e-free X e arc lam p (S p ectral E n ergy C o rp., W estw o o d N J) w ith a sp herical rear re¯ ecto r. A co nd en ser system con sistin g o f tw o plan o-co nv ex q u artz len ses is used to cou p le ligh t in to a m o n och rom ato r. T he p rim ar y con d enser is 1 .5 in . in diam eter w ith an ap ertu re ratio of f /1.5. T h e seco n dar y con d enser is also 1.5 in . in diam eter, b u t is m ask ed to pro vid e nu m erical ap ertu re m atch in g to the m o no ch ro m ato r. A m an u al shu tter is lo cated betw een th e co n d en sin g o p tics an d m on o chro m ato r an d is closed to p reven t ¯ uo rescen ce excitatio n lig ht from reach in g th e sam p le du ring re¯ ectan ce m easu rem ents. T h e m o no ch ro m ato r h as an aperture ratio o f f /3 .6 (S p ectral E n erg y, G M 2 5 2 ) an d is u sed w ith an io n-etched ho lo graph ic g ratin g (IS A , E diso n , N J; 2 4 0 n m b laze, 11 8 0 gro ov es/m m , d isp ersio n 5 3 .3 nm /m m ). A n R S -2 32

304


con tro lled step p er m o to r drives th e m o no ch ro m ato r w ith a m ax im u m step p ing rate of 40 0 step /s (app rox im ately 10 n m /s). A b and w idth o f 6 .6 n m is selected b y settin g th e en tran ce slit o f th e m on o chro m ato r to 2 .0 m m . L igh t is cou p led fro m th e m o no ch ro m ato r in to the p rob e v ia a ® b er-o ptic ad ap ter (S p ectral E n erg y, G M A 2 5 7) co n sistin g o f a q uartz plan o-co nv ex len s an d a 5 3 qu artz m icro scop e o b jective. T he ligh t passin g th rou g h th e o b jectiv e is focu sed on to a v ertical line o f 2 5 ® b ers in tw o colu m n s, placed at th e fo cal plan e o f th e o b jectiv e. T he re¯ ectan ce excitatio n ® b er is attach ed to the lam p h o usin g via a m icro p osition er. B ro ad ban d lig h t exiting the lam p ho u sing th ro u g h an ex istin g ho le is co u pled to the re¯ ectan ce illu m in atio n ® b er b y usin g a qu artz p lan o con v ex len s (N A 5 0 .2 4 ). A ® v e-po sitio n illum in ation ® lter w h eel p laced b etw een th e lam p an d th e len s con tain s th ree lo ng -pass ® lters w ith 5 0 % transm ission at 2 9 5 , 51 5 , and 7 15 nm . O n e o f th e ® lter p o sitio ns is blo cked and acts as a sh u tter to p rev en t w h ite lig ht fro m reach in g th e sam p le du rin g ¯ uo rescen ce m easu rem ents. L ig h t co llected b y ¯ u o rescen ce an d re¯ ectan ce ® b ers is co up led thro u gh an eig ht-po sition , co m p u ter-con tro lled collectio n ® lter w h eel in to a C h ro m ex 25 0 IS (A lbu qu erqu e, N M ) im ag in g spectro g rap h co n tain ing a h o lo grap hic g ratin g b lazed at 3 80 n m w ith 1 50 gro ov es/m m and a recipro cal linear disp ersio n (R L D ) of 20 n m /m m . T h e ® b ers are pro jected on to an en tran ce slit (2 50 m m ) th at y ield ed a spectral resolu tio n o f 5 n m . A th erm o electrically co o led C C D cam era op erated at 2 3 0 8 C (S pectrasou rce H P C -1 , W estlak e V illag e, C A ) w as lo cated at th e b ack fo cal p lane o f th e im ag ing sp ectro grap h. C h ip dim ensio ns w ere 1 3 .8 3 9 .2 m m w ith 15 3 6 3 10 2 4 pix els (K o d ak K A F -1 60 0 g rad e 2 ), yieldin g a no m in al spectral ran g e o f 27 6 n m fo r a sin gle gratin g p o sitio n. D ark cu rrent is sp eci® ed as 0 .2 5 electron s/pix el/s w h en op erated at 2 3 0 8 C . T he q u antu m ef® cien cy of the lu m o g en co ated ch ip rang es fro m a p eak o f 40 % at 5 5 0 nm to a lo w o f 1 5 % at 2 50 nm . T h e detector an d im agin g sp ectro grap h w ere w av elen g th calibrated b y m easuring the ro o m lig ht sp ectra th at sho w ed three m ercur y p eaks at 4 04 .7 , 4 3 6 , and 5 46 nm . T h e relation b etw een p ix els and w av eleng th w as th en linearly ® tted thro ug h th ese p oin ts. F lu o rescen ce an d re¯ ectan ce m easu rem en ts are ob tain ed seq u en tially. P rio r to ¯ uo rescen ce m easu rem en ts, th e w h ite lig h t p ort is clo sed and pix els illu m inated by th e ¯ u orescen ce ® b ers are selected to be read fro m th e C C D . D ark cu rren t and A /D con v ersio n o ffset is m easured w ith th e sam e settin g as th e sub seq uen t m easu rem en t bu t w ith a clo sed cam era sh u tter. T hese are sub tracted fro m all ¯ u orescence an d re¯ ectan ce m easu rem en ts. T h e ® rst ex citation w av eleng th is selected b y scan n ing th e ex citatio n m on o chro m ator, the em issio n ® lter w h eel is ro tated to select the ap p ro p riate lo ng -pass ® lter, an d th e spectro g rap h gratin g is ad justed to reco rd sig n al o v er th e d esired em ission w av elen gth ran g e. T he m o n och ro m ato r and cam era sh u tters are then op en ed fo r th e desired ex po su re tim e to reco rd th e ¯ uo rescen ce em ission sp ectru m (1.5 s). T he ex citatio n w av elen g th is th en in crem ented , an d th e p rocess rep eated u ntil all d esired ex citatio n w avelen gth s h av e b een m easu red . T h e excitatio n w avelen gth s w ere in crem en ted fro m 3 3 0 to 50 0 nm in 1 0 n m step s. Ta b le I co n tain s a list o f th e

T A B LE I. F luorescen ce lo n g- p ass a n d d iffu se re¯ ecta n ce ® lters u se d . F lu orescen ce lon g -p ass ® lters are sh o w n w ith th e corresp on d in g ¯ u oresc en ce ex cita t ion w av elen gt h . D iffu se re¯ ect an ce ® lters are sh ow n w ith th e corresp on d ing m ea su rem en t ran ge. F luor escen ce lo n g-p ass ® lter (1 ) (2 ) (3 ) (4 ) (5 ) (6 ) (7 ) (8 )

GG GG GG GG GG OG OG OG

375 395 420 445 475 495 515 530

3 30, 3 50, 3 70, 4 00, 4 30, 4 50, 4 70, 4 90,

R e¯ ectan ce ® lter (1 ) W G 29 5 1 (2 ) O G 515 1 (3 ) R G 7 15 1

F luoresc en ce excitatio n w aveleng ths (nm )

G G 37 5 G G 3 75 G G 3 75

3 40 3 60 3 80, 3 90 4 10, 4 20 4 40 4 60 4 80 5 00

P ossib le re¯ ectan ce m easu rem ent rang e 3 50 ±62 8 5 00 ±77 8 7 00 ±97 8

ex citatio n w av elen g ths and corresp o n din g lon g -p ass ® lters an d em issio n w avelen gth rang es used in this stud y. F ollow in g co llectio n of ¯ u o rescen ce sp ectra, d iffu se re¯ ectan ce spectra are th en m easured . F or these m easu rem en ts, th e m on o chro m ato r sh utter is clo sed , th e em issio n ® lter w heel is set to th e lo w est ® lter p osition , an d th e p ix els illu m in ated by th e co rresp o nd ing re¯ ectance collectio n ® b ers are selected to be read fro m the C C D . D ark cu rrent an d A /D co nv ersio n o ffsets are m easured an d sto red for su b traction o f th e follo w in g m easu rem en ts. T h e re¯ ectan ce sp ectru m is co llected o ver three illu m in atio n w av elen g th ran g es. P rior to m easu rem ent o f each rang e, th e ap p ro p riate lo n g-p ass ® lter is selected in th e illu m in atio n ® lter w heel, an d the sp ectrog raph g ratin g is ad ju sted to reco rd sig nal o v er th e desired w avelen gth rang e. T he lam p an d cam era shu tters are then o p en ed fo r the desired ex po su re tim e to record th e re¯ ectan ce sp ectru m (0 .4 ±4 .8 s). T h e illum in ation w av eleng th ran g e is then increm en ted , an d the pro cess rep eated u n til all d esired w avelen gth ran g es h av e b een m easu red. E x po su re tim es are d eterm ined em p irically to achiev e an S N R g reater than 20 . Ta b le I con tains a list o f the illu m in atio n w avelen gth rang es an d correspo n din g lo n g-p ass ® lters u sed fo r d iffu se re¯ ectance m easu rem ents. T h e hig h d y n am ic rang e o f the re¯ ectan ce m easurem en ts, span n in g o ver three ord ers o f m agn itud e, requ ires th at each sp atial p osition b e read ou t ind ivid u ally fro m th e C C D . T h is ap pro ach preven ts saturatio n an d b lo om ing artifacts. T h ere are n o accep ted safety stan dard s fo r illu m in atio n o f m u co sal sur faces oth er th an sk in and co rn ea. H o w ev er, w e calcu lated th e exp o sure o f solar rad iatio n that is eq uiv alen t to th e ex p osu re receiv ed w h en a m easu rem en t is m ad e w ith ou r sy stem . T h e m etho d co m p ares the sp ectral irradian ce (W /cm 2 n m ) of th e excitatio n sou rce w ith so lar irradian ce d ata o b tain ed fro m R ef. 4 0 . T h e com p ariso n con sists of a p oin t-w ise d iv isio n o f th e irradian ce fro m th e F astE E M sy stem to th e solar irrad iance at th e sam e w av eleng th. T h is ratio g iv es a relativ e so lar ex p o su re facto r. T he so lar data are fo r a su n ny d ay in S an D ieg o , C A . Irrad iatio n du rin g ¯ u o rescen ce ex citation is less th an 7 tim es so lar ex p o sure at all w avelen gth s. G iv en that ¯ u o rescen ce excitatio n tim es w ere 1 .5 s, this co rresp on d s to ex po su re to so lar radiatio n fo r less th an 1 1 s

in an y g iv en w av eleng th ban d . D u rin g d iffuse re¯ ectance m easu rem en ts, the lam p ex po su re is m ax im u m at 3 0 0 n m , w h ere th e relativ e ex p osure is a facto r o f 2 5 that o f the su n . S in ce the to tal ex po su re tim e fo r this w av eleng th b and is 14 s, th e ex p osure co rresp on d s to 3 50 s o r less than 6 m in . A ll oth er w av elen g th s h av e relativ e exp o sure facto rs o f 1 0 o r less, resu ltin g in a sho rter eq uiv alent to tal so lar ex p osu re. P rio r to ev er y p atien t m easu rem en t, th e p ro b e o utp u t w as m easu red w ith a calibrated p ow er m eter (N ew p o rt, Ir v in e, C A , 8 1 8 -U V ) at 40 0 n m ex citatio n w av eleng th . A n av erag e ou tp ut of 8 6 m W 6 1 2 m W w as achiev ed at this w av elen g th w ith a b an d w id th of 6 .6 nm . B ackg rou n d ¯ uo rescen ce sp ectra w ere m easured w ith the p ro b e d ip p ed in a no n ¯ u o rescen t b o ttle co n tain in g d istilled w ater. T h is back gro un d E E M w as su btracted fro m all sub seq uen tly acq u ired E E M s to co rrect fo r ro om ligh ts an d p ro b e auto ¯ u o rescen ce. T h e n on u nifo rm spectral resp on se of th e sy stem w as co rrected b y u sin g co rrectio n facto rs d eterm in ed fro m m easurem en ts of calibratio n so urces; in th e visib le, w e u sed a N atio n al In stitu te o f S tan d ard s an d Te ch n o lo g y (N IS T ) traceable tu n gsten rib b on ® lam en t lam p , and in th e U V, w e u sed a d eu teriu m lam p (5 5 0 C an d 4 5D , O p tro nic L abo rato ries In c., O rlan d o, F L ). Va riatio n s in th e in ten sity of th e ¯ uo rescen ce ex citatio n lig h t sou rce at different ex citation w av elen g th s w ere corrected by usin g m easurem en ts o f th e in ten sity at each ex citatio n w av elen g th at th e p rob e tip w ith th e u se o f a calib rated p ho to dio d e (8 1 8 -U V, N ew p o rt). B ack g ro u nd sp ectra to co rrect re¯ ectance m easurem en ts fo r roo m ligh t co n trib utio ns w ere m easured w ith all p aram eters set as fo r tissu e m easu rem en ts, ex cep t th at th e w hite ligh t sh u tter w as clo sed. T h ese m easurem en ts w ere su b tracted from all su b sequ en t re¯ ectan ce sp ectra. F luo rescen ce an d re¯ ectance stan dards w ere m easured b efo re each patient m easu rem en t. T h e ¯ uo rescen ce in tensity is rep o rted relativ e to th e ¯ u orescen ce inten sity o f a so lution o f 2 m g /L R h o dam ine 6 1 0 (E xciton , D ayto n, O H ) in eth y len e gly col at 46 0 n m ex citation an d 5 8 0 n m em issio n. R e¯ ectance d ata are rep o rted w ith resp ect to a 2 .6 8% -b y -v o lum e so lu tio n o f 1 .0 72 m m d iam eter po ly sty ren e m icro sp heres (P o ly scien ce In c., W arrin g to n , PA ). T h e m icro sp here stan d ard w as used for its w ell-characterized op tical p ro p erties. T he total in teg rated re¯ ectance o f th is stan d ard w as m easured o n a d o ub le-b eam sp ectro p h oto m eter (U -3 3 00 H itach i, To k y o, Japan ) w ith an integ ratin g sph ere attachm ent (L absp h ere In c., N o rth S u tton , N H ). T h is m easu rem ent w as u sed to co rrect the re¯ ectan ce stan d ard m easu rem en ts m ad e w ith th e F astE E M sy stem . T issu e sp ectra at each co llection ® ber po sitio n w ere d iv id ed p o in tw ise b y th e corrected stan dard re¯ ectance spectru m at th e corresp o n din g ® b er po sitio n . T h e E E M s w ere assem bled o f¯ ine from each series of ¯ uo rescen ce em issio n scan s. D ata p ro cessing an d p lo ttin g w ere p er fo rm ed w ith M atlab (T he M ath W o rk s Inc., N atick , M A ). R e¯ ectan ce spectra w ere assem b led fro m th ree w avelen gth areas, g ivin g a rang e fro m 38 0 to 9 5 0 nm . T h e w av elen gth rang e w as fu rther redu ced (3 8 0 ±8 00 n m ) to co m p ly w ith th e ran g e of calib ratio n m easurem en ts o f the re¯ ectan ce stan dard s on the U -3 3 00 . R e¯ ectance d ata are rep o rted betw een 38 0 an d 5 9 5 n m , a ran g e w h ere the p ossib le in ¯ u ence of ro om ligh ts in th e m easu rem en t w as m in im ized.


305

S ystem V a lid ation . T h e sy stem p erfo rm an ce w as assessed b y u sing tw o ¯ u o rescen ce stand ards. T h e ® rst stan d ard w as a 2 m g /L R h o dam in e 6 1 0 (E xciton Inc., D ay ton , O H ) ethy lene g lyco l so lu tion , w hich is n on scattering b u t has p eak ¯ u o rescen ce in ten sity ap pro x im ately tw ice the av erag e inten sity o f h u m an cer v ix. 41 T he seco n d stan d ard m im ics th e o ptical pro perties of tissu e an d co nsists of 2 0 m M F lavin A d enin e D in u cleo tide (FA D , K o dak , R o ch ester, N Y ), 0 .6 25 v o l % p o ly sty ren e m icro sph eres (P o ly science In c., d iam eter 5 1.0 72 m m ). 4 2 B o th stan dard s w ere m easured w ith th e F astE E M system an d a scann in g sp ectro ¯ u orim eter (S P E X , F luo rolo g II, E d ison , N J). T he E E M s m easu red w ith th e S P E X w ere con sidered as stan dard s sin ce the p erfo rm ance o f the sy stem is w ell d o cum ented (d y nam ic ran g e 5 1 0 5 , sp ectral reso lutio n 5 n m , co rrected fo r no n un ifo rm spectral respo n se). T h e excitatio n lig h t w as incid ent p erpen d icu lar to th e sam p lin g cu v ette, an d th e em itted lig ht w as co llected at ap pro xim ately a 2 0 8 an gle w ith resp ect to th e excitatio n lig ht. A fro nt fo cu s arrang em en t w ith a 1 0 m m cuv ette w as u sed in th e S P E X . S ix ty m inu tes w as requ ired to co llect a fu ll E E M fro m each sam p le w ith th e SPEX . C lin ical S tu d ies. In vivo data w ere o b tain ed fro m a gro up o f p atien ts w ith k n ow n o r su spected p rem alig nan t or m alig n ant lesio ns of th e o ral cav ity. T h e stud ies w ere rev iew ed an d app rov ed b y th e In ternal R ev iew B o ard of th e U n iv ersity o f Te xas at A u stin and th e S u r v eillan ce C o m m ittee at th e U T M . D . A n d erso n C an cer C enter (H o usto n ). Info rm ed co nsen t w as ob tained fro m each p erson in th e stu d y. B efo re th e pro be w as used , it w as d isin fected w ith M etricid e (M etrex R esearch C orp .) in accord an ce w ith the stan d ard clin ical p ro to co l. B ack gro un d ¯ u o rescen ce E E M and re¯ ectance spectra w ere m easu red by d ip pin g th e ® b er-o ptic pro be in a n on ¯ uo rescen t b o ttle ® lled w ith d eion ized w ater. T hese E E M s an d spectra co rresp o nd to th e sy stem auto ¯ u o rescen ce an d w ere su b tracted fro m all su b seq uen tly acq u ired E E M s fo r th at p atien t. N ext an E E M w as m easu red from a R h od am in e calib ration stan dard , and a re¯ ectan ce sp ectru m w as m easured fro m a p oly styren e so lu tion calib ratio n stan d ard . T h e p ro b e w as th en g uid ed to th e tissu e site to be ex am ined an d its tip p o sitio ned ¯ u sh w ith the tissu e. A ¯ u o rescen ce E E M an d re¯ ectance spectra w ere o b tain ed fro m sites w ith in a lesion an d a clin ically no rm al site. P o st-sp ectro sco py, a 2 ±4 m m b iop sy of th e tissu e w as tak en from no rm al an d ab no rm al sites w h ere th e p ro b e m easu red sp ectra. T hese sp ecim ens w ere ev aluated by an exp erien ced path o lo g ist, B o nn ie K em p, M .D ., w ith the use o f lig h t m icro scop y an d w ere classi® ed b y usin g standard d iagn o stic criteria. D ata A n a ly sis. O n e of the g oals of instru m en t d ev elop m en t is to p rov ide in fo rm ation fo r th e id enti® cation o f excitatio n w avelen gth s su itab le for th e differen tiatio n o f tissu e o f d ifferin g path olo g ical characteristics, as w ell as id en ti® catio n o f th e ch ro m op h ores resp o n sib le fo r th e differences. W h ile all su ch in fo rm ation is p resent in the E E M s co llected , it can be d if® cu lt to ex tract d u e to th e dim en sio nality of th e d ata set. A m eth o d w as d evised to separately ch aracterize th e excitatio n an d em ission characteristics of th e data set. G iv en th at th e E E M has dim en sio n s correspo n din g to ( l x , l m ), th e fo llo w in g au to correlation vecto rs are d e® n ed :

306


x av ( l m av ( l

x

) 5

m

) 5

O

N

O

i5 1

E E M (l

x

, l

E E M (l

xi

mi

) ´ E E M (l

x

, l

mi

)

m

) ´ E E M (l

xi

, l

m

)

N i5 1

, l

w h ere x av ( l x ) is th e ex citatio n au to co rrelatio n v ector an d m av ( l m ) is the em issio n au to co rrelatio n v ecto r. E ssen tially, th e em issio n auto correlation v ecto r is th e diag on al o f th e pro du ct o f th e E E M w ith its transp o se, and th e ex citation auto correlation v ecto r is th e d iag o nal of the pro du ct o f th e tran sp ose o f th e E E M w ith the E E M . N o te that in sig n al p ro cessin g term s, th e au toco rrelatio n vecto rs, x av and m av , are a m easu re o f the av erag e sig nal o f th e E E M at each ex citation o r em issio n w avelen gth , respectively. In th is w ay they p rov id e qu alitative info rm atio n ab o ut an EEM . A n ex am p le w ith sim u lated d ata is presented in F ig . 3 to illustrate h o w au toco rrelatio n vecto rs re¯ ect ch ang es in ¯ u orescence p eak p o sitio ns in E E M s. Tw o kin d s of chan g es are sim ulated in th e m o deled d ata: a shift in the excitatio n w avelen gth at w h ich a ¯ uo rescen ce peak appears, an d a shift in the em issio n w av eleng th at w h ich a ¯ u o rescen ce peak ap pears. T h e o rig inal p eak in th e E E M w as m od eled as a sin g le G aussian at 3 80 nm ex citation , 55 0 n m em ission w ith a fu ll w idth at half-m ax im u m (F W H M ) of 3 5 n m in em issio n an d ex citatio n w av elen g th s. T h e o rig in al peak w as th en sh ifted by 3 0 n m in excitatio n, as sh ow n b y arro w 1 in F ig. 3 a. T h e sh ift in em ission w av elen g th is sh ow n by arro w 2 in F ig . 3a, an d corresp o n ds to a 30 nm shift in th e em issio n p eak of th e origin al d ata. T h ree sets of auto co rrelatio n vectors w ere com pu ted: on e fo r the E E M w ith th e origin al p eak , o n e fo r th e E E M w ith th e ex citatio n w aveleng th -sh ifted p eak , and on e fo r the E E M w ith the em issio n w av eleng th-sh ifted peak . T h e auto co rrelatio n v ectors are sh ow n in F ig . 3b . C o m p arin g th e v ecto rs fo r the o rigin al E E M (row 1 in F ig . 3b ) w ith the vecto rs fro m th e E E M w ith th e ex citatio n w av elen g th -sh ifted E E M (row 2 in F ig . 3 b), it is seen th at th e excitatio n au to correlation vecto r is sen sitiv e to th e ch ang e in ex citation w av elen g th b u t n o t in em issio n w av eleng th. S im ilarly, com p arin g th e auto correlatio n v ecto rs fo r the orig in al E E M w ith the vecto rs fro m th e E E M w ith th e em ission w av elen g th sh ift in th e p eak (ro w 3 in F ig. 3 b ) sho w s that th e em ission au to co rrelatio n vecto r is sen sitiv e to th e ch ang es in em issio n w av eleng th bu t no t ex citatio n w avelen gth . It is so m etim es d esirab le to n orm alize the au to co rrelatio n v ecto rs to facilitate com parison s betw een d ifferen t sets o f m easu rem ents. W e calcu lated no rm alized au to correlation v ecto rs b y d iv id ing these v ectors b y their ro ot m ean sq u are (rm s) v alue, in effect fo rcin g th e area of th e vecto r to 1 u nit o f sig nal en erg y. T h e no rm alized em issio n au to co rrelatio n v ector is w ell suited for th e id enti® catio n o f d ifferen tial featu res in E E M s, such as th e sh iftin g or bro aden ing o f ¯ u orescence peak s. R E S U L T S A N D D IS C U S S IO N F ig u res 4 A an d 4 B sho w ¯ u orescence E E M s o f th e no n scatterin g R h od am in e stand ard an d th e scattering FA D ph an to m o b tain ed w ith th e F astE E M sy stem . In ten sities are rep orted w ith resp ect to th e R h o dam in e in ten -

F IG . 3. (a ) S im ulated E E M w ith peak sh ifting in (1) excitation w avelength and (2 ) em issio n w avelength. ( b ) C alculated x av and m a v f or the sim ulated E E M . x av is se n sitive to chang es in the ex citation po si tion of the peak, an d m a v is se n si tive to the em issio n positio n .

sity m easu red at 4 60 nm ex citation an d 58 0 n m em issio n w avelen gth . F ig u res 5 a an d 5c sh o w ¯ u orescence em issio n sp ectra o f th e R h o dam in e stan d ard o b tain ed at 3 7 0 an d 4 5 0 n m ex citatio n w ith th e S P E X an d th e F astE E M sy stem as w ell as th e ¯ uo rescen ce b ack gro un d . F ig u res 5 b an d 5 d sho w th e sam e sp ectra fo r scatterin g FA D p han tom o btained at th e sam e ex citatio n w av eleng th s. T h e spectra are n orm alized at th eir m ax im u m . N o te the p resen ce of R ay leigh scattering p eak s fro m th e excitatio n so urce in th e d ata tak en w ith the S P E X . In gen eral, fro m n on scattering sam p les (F ig s. 5a, 5 c) th e F astE E M sy stem co llects less lig h t ab o ve 6 0 0 n m th an th e S P E X . T h is result co u ld b e du e to th e d ifferen t co llectio n ef® cien cies o f the F astE E M pro be an d th e fro nt-face collectio n geo m etr y o f th e S P E X . U nd er scatterin g co nd itio n s and w ith low er ¯ u o rescen ce sig nal, th e in ¯ u ence o f b ackg ro u nd ¯ u o rescen ce beco m es m o re critical. A t 3 70 n m

F IG . 4. (A ) E E M of R hodam in e stan dar d so lution . (B ) E E M of an FA D an d m icr osp heres-b ased tissu e ph an tom m easu red w ith the u se o f the F astE E M sy stem .

ex citatio n w avelen gth , the F astE E M system m easu res m o re ¯ u o rescen ce b elow 50 0 nm . A co m p ariso n w ith th e m easu red ¯ uo rescence b ack g ro u n d, h o w ev er, sho w s that the ad d itio nal sig nal h as the sam e sh ap e as th e back g ro u nd . W e h y po thesize th at the b ack g ro u nd has b een u nd erestim ated by m easu rin g it in a no n scattering n o n ¯ uo rescen t m ed ia. In vivo ¯ uo rescence E E M s o f th e o ral cav ity w ere m easu red fro m 7 1 sites, an d in vivo re¯ ectance sp ectra w ere m easured from 4 9 sites. T h ese w ere ob tain ed fro m p atien ts in tw o stud ies. T he ® rst stu d y in v olv ed p atien ts w ith abn o rm al oral lesio ns iden ti® ed in a p revio u s m edical exam ination (1 7 p atien ts). T he seco n d stu dy, in clud ing nin e p atien ts, in v o lv ed n orm al vo lun teers. A ll sites interro g ated sp ectro scop ically in patients w ith lesion s w ere b iop sied and su b m itted for h isto p ath o lo g ical analy sis. S p ectra an d bio p sies w ere also ob tained fro m a con tralateral site w ith no lesio n in th ese p atien ts w ith ab n orm al lesion s. T h ese b iop sies w ere also ev alu ated h istop ath olo gically. N o b io p sies w ere tak en from th e no rm al


307

F IG . 5. (a) E m issi o n sp ectra at 3 60 nm excitation of the Rho dam ine calibr ation stand ar d m easu r ed w ith the F astE E M sy stem and S P E X F lu o ro lo g II ¯ u orim eter. (b ) E m issio n sp ectra at 3 60 nm excitation o f th e sc attering tissu e p h anthom con taining FA D and polysty r en e m icrosp heres m easu red w ith the F astE E M sy stem and S P E X F luo r o log II ¯ uo rim eter. (c) E m issio n sp ectra at 4 50 n m excitation o f th e Rh o dam ine calibratio n stan dar d m easu r ed w ith the F astE E M sy stem an d S P E X F luo ro log I I ¯ u or im eter. (d ) E m issio n sp ectra at 450 nm excitation of the sc attering tissu e p hantho m containing FA D an d po lyst yrene m icr osp heres m easu red w ith the F astE E M sy stem and S P E X F luo ro log II ¯ uo r im eter.

vo lu nteers. H ere, w e sho w rep resen tativ e E E M s fro m tissue fo un d to b e h isto p ath o lo g ically n o rm al and m alig nan t to illustrate sp ectral features d etectab le w ith th e F astE E M system . Tw o E E M con tou r plo ts fro m a n o rm al an d an abn o rm al area of the to ng u e are p resen ted in F igs. 6 A an d 6 B , resp ectiv ely. In th e n o rm al sam p le, ¯ uo rescen ce is ob ser v ed th ro u g ho u t th e w h ole collectio n ran g e, w ith a peak lo cated at 33 0 /38 0 (ex citatio n/em issio n) an d a rid g e exten din g fro m 3 4 0/4 5 0 to 4 5 0/5 0 0 nm . Ta b le II lists excitatio n-em issio n m axim a p airs o f en do g eno u s tissue chro m o ph o res. C om p ariso n o f the o b ser v ed p eak s w ith Ta b le II sho w s that these p eak s are co nsisten t w ith the em ission o f structu ral pro tein s su ch as co llag en an d elastin , p y rid ine n ucleo tid es (N A D H ), and ¯ avo p roteins (FA D ). T h e n o rm al site sh o w s o verall increased ¯ u o rescen ce w ith resp ect to the ab n o rm al site sh ow n in F ig . 6 B . T h e abn o rm al site, assessed b y a p ath o log ist as b eing m o d erately d ifferen tiated sq uam o us cell carcin om a, also sho w s bro ad ¯ uo rescen ce th rou g h ou t. P eaks are ob ser ved at 3 30 /3 80 , 3 5 0/4 6 0, 46 0 /52 0 , and 5 00 /63 0 n m . A v alley is seen at 42 0 n m ex citation b etw een 56 0 and 58 0 n m em ission . T his v alley is seen to ex ten d alo n g th e 4 20 n m excitatio n lin e as w ell as th e 5 8 0 n m em issio n lin e. Ta ble III su gg ests that these featu res are p ro d u ced by h em o g lobin reabso rptio n. H em o glo bin reab so rp tion m ay also in part acco u n t fo r th e shift in th e peak s of the ab n o rm al E E M w ith resp ect to th e n orm al E E M . A su m m ar y of th e ex citatio n and em issio n m ax im a fo r th e p eak s o bser v ed in th e n orm al an d abn o rm al sites m easured is p resented in Ta ble IV .

308


F lu o rescen ce em issio n sp ectra at three selected excitatio n w avelen gth s are sho w n in F ig . 7, illu stratin g chan g es in relative in ten sities o f ¯ u orescence em issio n. F o r co m p ariso n p u rp o ses, each set (n o rm al/ab n orm al) w as n o rm alized to th e m ax im u m at 35 0 nm excitatio n . F ig ure 7 a sho w s th e em issio n sp ectra at 35 0 n m ex citatio n . F luo rescen ce from th e no rm al site is seen as a b ro ad peak w ith a m ax im u m at 45 5 nm . T he p eak fro m the abn o rm al site is seen to be n arro w er an d red -sh ifted . E xam inatio n o f th is spectru m at 41 0 , 5 40 , and 5 80 nm su g gests th at th e chan g e in line sh ap e is d ue to o xy g enated hem og lo bin . T h e g en eral lin e sh ap es o f th e ¯ u o rescen ce ob ser ved at 41 0 n m ex citation (F ig. 7 b) are seen to b e sim ilar fo r bo th sites in the 4 50 ±5 7 5 n m em ission rang e, w ith a bro ad peak at 5 0 0 nm . T he ab n orm al site sh ow s a sig ni® can tly low er ¯ u o rescen ce in tensity, as w ell as an extra, n arrow ¯ u orescence p eak at 64 0 n m , attrib uted to po rph y rin ¯ u o rescen ce. F ig ure 7 c sho w s th e em ission spectra at 4 6 0 n m ex citation . T he no rm al site sh o w s a bro ad p eak at 52 0 nm an d clear m od u latio n fro m heTA B L E II. S u m m ary of ex citat ion /em issio n m axim a o f en d o g en ou s t issu e ch rom op h ores. C hr om oph ores NADH FA D FA D P or phy rin Co llagen I Co llagen I E lastin

E x citation /em issi on ( nm ) 34 0/450 1 1 45 0/515 1 1 37 0/535 1 1 42 0/640 1 1 34 0/410 4 3 50 0/520 4 3 33 0/405 4 3

T A B LE IV . S u m m ary of excitation /em issio n m ax im a f o r p eak s in th e ¯ u o resc en ce E EM s of n orm al an d a b n orm al oral cavity sa m p les. T h e in ten sity is rep orted rela tive t o t h e rh od am in e st an d a rd at 46 0/ 580 n m . P eak location [ E x citation/em issio n nm ]

F luor esc ence intensity (R elative to rh odam ine stan dard at 460/58 0 nm )

N or m al site ( F ig. 6A ) 330/38 0 340/45 0 450/50 0

0 .5 97 0 .5 83 0 .2 38

A bnor m al site (F ig. 6B ) 330/38 0 350/46 0 460/52 0 500/63 0

0 .0 82 0 .1 66 0 .0 31 0 .0 32

E E M s. T h e v ectors sh o w th e effect o f h em o g lo b in ab so rp tion arou n d 41 0 , 5 40 , and 5 80 n m in the ab n orm al site an d the p resence of add itio n al ¯ u o rescen ce in th e U V in th e n o rm al sam p le (F ig. 8 A ). T his au to co rrelatio n v ector also h igh ligh ts the p eak at 6 10 n m in th e ab no rm al sam p le. T h e excitatio n au to co rrelatio n v ectors sh o w d ifferen t line sh apes. T h e cu r ve correspo n din g to th e n orm al site d ecreases stead ily fro m 33 0 to 5 0 0 n m ex citatio n. T h e cu r v e fro m th e ab no rm al site sh o w s a p eak at 3 5 0 an d a m in im u m at 4 10 n m . T h e latter illu strates the g reat-

F IG . 6. In vivo ¯ uo resc en ce m easu rem ents w ith th e F astE E M sy stem : (A ) F luoresc ence E E M of a nor m al site of th e to ngue. (B ) F luoresc en ce E E M of a dise ased site of the tong ue, con taining a m o derately differentiated sq u am ou s cell carcino m a.

m o g lob in reabso rptio n at 54 0 an d 5 8 0 nm . F lu o rescen ce fro m th e abn o rm al site sho w s an ev en m ore m ark ed h em o g lob in reab sorp tio n ; also the o verall ¯ u orescence in tensity is redu ced . F igu re 8 sho w s th e em issio n and ex citatio n au to co rrelation vecto rs fo r the sam e m easu rem ents. N o te th at the p lo ts h ave a lo garith m ic y axis. T he em issio n au to co rrelation vecto rs h av e a larg e b road p eak at 4 60 n m co rrespo n din g to th e m ain ¯ u orescence p eak o b ser v ed in th e T A B LE III. S u m m ary of a b so rp tion m ax im a of ox y- an d d eoxyh em oglob in. A bso rb er s

A bso rp tion w avelength ( nm )

O xygenated h em og lob in O xygenated h em og lob in O xygenated h em og lob in D eoxygenated hem og lob in D eoxygenated hem og lob in

4 15 4 4 5 42 4 4 5 77 4 4 4 30 4 4 5 55 4 4

F IG . 7. F luoresc en ce em issi o n sp ectra of no rm al and m o derately d iffer entiated sq u am ous cell carcinom a of the to ngue f ro m F ig. 6 . T h e sp ectra w ere nor m alized to th e peak ¯ uoresc ence at 3 50 nm ex citation . (a) F luor escence em issio n sp ectra at 35 0 nm ex citation. ( b ) F luo r escence em issio n sp ectra at 410 nm excitation . ( c) F luoresc en ce em issio n sp ectra at 4 60 nm excitation.


309

F IG . 8. E m issio n an d ex citatio n au tocor relation vector s of nor m al and m oder ately d ifferen tiated sq uam o u s cell carcino m a of the tong ue f ro m F ig. 6. (a) E m issio n au tocorr elation vectors. (b ) E xcitatio n auto co rr elation vectors.

er in ¯ u en ce of h em o g lo b in reab sorp tio n in th e ab n orm al sam p le, also sho w n in F ig . 7 . T h e correspo n din g re¯ ectan ce d ata are sh ow n in F ig . 9. P o sitio n 1 co rresp o nd s to th e co llectio n ® b ers closest to the sou rce ® ber an d p o sitio n 3 to th o se furthest fro m th e so urce ® b er, as sho w n in F ig . 2 . D ifferen ces in d u ced by the ¯ uo rescen ce reab sorp tio n o f o x yg en ated hem og lo bin in the no rm al site an d ab n orm al site are sh o w n . T h e m o d ulatio n o f th e sp ectru m b y th e 5 40 and 5 8 0 n m ab sorp tio n ban d s is seen to b e sig n i® can tly stro n ger in the abn o rm al sam p le; this o bser v atio n is con sistent w ith th e in creased reabso rptio n seen in the ¯ u o rescen ce sp ectrum of th e ab n orm al sam p le. T h e re¯ ectan ce in th e b lue ran ge (4 5 0 ±50 0 n m ) o f th e ab n orm al site is co n sisten tly hig her th an th at of th e n orm al site. B elo w 4 5 0 nm th e re¯ ectan ce seem s no t to d iffer b etw een th e no rm al an d abn o rm al sam p les. C O N C L U S IO N T h e total data acq u isition tim e for th e d ata presen ted here w as 2 .5 m in fo r a ¯ u o rescen ce E E M an d 1 .5 m in fo r the sp atially reso lved re¯ ectan ce m easu rem ents. H o w ever, o nly 2 9 s o f th is tim e represen ts ¯ u orescence co llection . A ctual re¯ ectan ce co llectio n tim e w as 26 s. T he m o st tim e-co n su m ing pro cess w as ch an gin g th e excitatio n w avelen gth b y usin g the step per m o to r-co n tro lled ex citatio n spectrog raph and chan g in g th e co rresp o nd in g lo n g -p ass ® lter w ith the use of th e rem o tely co ntro lled ® lter w h eel. W o rm d riv e-b ased m o n o ch rom ators are available (D D D 1 80 , IS A ) that requ ire less th an 10 s to scan ou r en tire w avelen gth rang e in 10 n m steps an d cou ld sub stantially red u ce the to tal m easu rem en t tim e. U sin g a h igh er po w er lam p w ill further red u ce acq u isition tim e o f b oth ¯ u o rescen ce an d re¯ ectan ce. W e hav e d em o nstrated th e acq uisition of E E M s in com bin atio n w ith sp atially resolv ed re¯ ectan ce m easurem en ts o f tissue ph an to m s an d in th e o ral cav ity in vivo w ith g o o d sig nal-to-n oise ratio. T h e sy stem featu res easy

310


F IG . 9. en tiated arations P osition

Re¯ ectance m easu rem en ts o f nor m al an d m o derately differsq uam o us cell car cino m a o f the ton gue at th ree d ifferen t sep f ro m the so u r ce ® b er. (a) P osition 1, 1.1 m m se p aration . (b ) 2 , 2.1 m m se par ation. (c ) P osition 3, 3 m m se paration .

and arbitrar y selection o f ex citatio n w av elen g th s in th e U V an d v isible ran ge. T h e sy stem is also p ortable and is cap ab le o f fun ctio n ing in a h ospital o peratin g ro o m . P rob es u sed in th e F E E M sy stem inco rpo rate chan n els to m easu re sp atially reso lv ed re¯ ectan ce and ¯ u o rescen ce and are m ad e sm all en o u gh ( , 5 m m ) to b e u sed d uring end o scop ic surg ical p ro cedu res. A u toco rrelatio n vecto rs x av and m av are a suitable m eth o d to red uce th e data set w h ile p reser v in g in fo rm atio n ab ou t th e w aveleng th b an ds carr yin g in fo rm atio n. O n th e b asis of th e representative data sh ow n h ere, ¯ u o rescen ce em issio n an d ex citatio n as w ell as re¯ ectan ce d ata ap p ear p rom isin g for th e id enti® catio n of tum o rs o f th e o ral cavity. T h e F astE E M system is an ideal to o l to id en tify a su bset o f th e m o st p ro m isin g o ptical features to iden tify path o lo g ical ® n d in g s in larg e clin ical stu dies. A CK N O W L E D G M E N T T his w ork w as su ppo rted by a gr ant f ro m the W hitaker F o u nd ation.

1 . N . J. S terenbor g, S . T hom se n, S . L . Ja cq ues, M . D uvic, M . M o tam edi, an d R. F. Wagner, Jr., D er m atolog ic S ur g er y 21 , 8 21 ( 19 95 ) . 2 . A . G illenw ater, pap er presen ted at I ntern ational C o nfer ence o n H ead and N eck C an cer, Tor o nto, C anada, ( 199 6) . 3 . D . R . I ngr am s, J. K . D hin g ra, K . R oy, D . F. P errault, Jr., I. D . B o ttrill, S . K abani, E . E . R ebeiz, M . M . P ankr atov, S . M . S h apsh ay,

4.

5. 6.

7.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

23.

R . M ano haran , I. I tzk an, an d M . S . F eld, H ead & N eck 19, 27 ( 1997) . N . R am an ujam , M . F. M itch ell, A . M ahadevan-Ja nse n, S . L . T hom se n, G . S taerk el, A . M alpica, T. Wr ight, N . A tk inso n, an d R . R ichar ds- K o r tum , P ho to ch em . P hotobiol. 64, 7 20 (1 9 9 6). F. K o enig , F. J. M cG o ver n, A . F. A lthau se n , T. F. D eutsch , and K . T. S cho m acker, J. U ro l. 156, 1 597 (1 9 96). K . T. S ch om acker, J. K . F r iso li, C. C . C o m pto n, T. J. F lotte, J. M . R ichter, T. F. D eu tsch , an d N . S . N ish iok a, G astro en terolog y 102, 11 55 ( 19 92 ) . R . M . C othren , R . R ichard s- K o rtu m , M . V. S ivak, M . F itzm aur ice, R . P. Rava, G . A . B o yce, M . D ox tader, R . Blackm an, T. B. I vanc, G . B . H ay es, M . S . F eld, an d R. E . P etras, G astro intest inal E ndo sc. 36 , 105 (1 99 0). Y. Ya n g , A . K atz, E . J. C elm er, M . Z uraw sk a-S zczepaniak, and R . R . A lfan o , L aser s L ife S ci. 7, 1 15 (1 9 96 ). Y. G . C h un g , J. A . S ch w artz, C . M . G ard n er, R . E . S aw ay a, an d S . L . Ja cq ues, J. K o rean M ed. S ci. 12, 13 5 (1 99 7 ). N . S . N ish iok a, G astr ointestinal E n dosc . C linics N o rth A m . 4, 313 ( 1994) . R . R ich ards-K or tum and E . S evick-M u raca, A nnu . R ev. Ph ys. Chem . 47 , 555 (1 99 6). A . G . B o ho r f o ush , E n d osc o py 28 , 372 (1 99 6). J. J. B igio an d J. R . M o urant, P hysics M ed . B iol. 42, 80 3 ( 1997) . J. F. H ollan d, R . E . Teets, and A . Tim m ick, A n al. Chem . 45, 145 ( 1973) . I . M . War n er, J. B. C allis, E . R . D avidsio n, M . G o uterm an , and G . D . C hristian , A nal. L ett. 8, 6 65 ( 19 75) . I . M . Warn er, G . D . C hristian, E . R . D avid so n , an d J. B . C allis, A n al. C h em . 49 , 56 4 ( 1977 ) . I . M . War n er, E . R . D avidso n, an d G . D . Ch ristian, A n al. C hem . 49 , 215 5 ( 19 77 ). C .- N . H o , G . D . C h ristian , an d E . R. D avidso n, A nal. C hem . 50, 11 08 ( 19 78 ) . B . E . W ilson , W. L ind ber g, and B . K ow alsk i, J. A m . C hem . S oc. 11 1, 37 9 7 ( 1 98 9) . A . J. D ur kin, S . Ja iku m ar, N . R am an ujam , and R . R ichard s-K o rtum , A p pl. O pt. 33 , 4 1 4 ( 1 994) . A . J. D u r k in and R. R ichard s-K o r tum , L ase rs S urg er y M ed. 19 , 75 ( 1996) . L . I . D eck elbau m , M . L . S tetz, K . M . O ’ B rien, F. W. C utru zzo la, A . F. G m itro, L . I. L aifer, and G . R . G indi, L aser s S u rg er y M ed . 9, 20 4 ( 198 9) . A . M ahad ev an, M . F. M itchell, E . S ilva, S . T hom se n, an d R . R ichar ds- K o r tum , L ase rs S ur ger y M ed. 1 3, 64 7 ( 1993 ).

24. R . R ichard s-K or tum , R . P. Rav a, R . E . P etras, M . F itzm au rice, M . S ivak, and M . S . F eld, P hotochem . P hotob iol. 53, 7 77 ( 1 99 1) . 25. R . Co thren, T. K olubayev, B . T. K jellstrom , R . R ichard s- K o rtu m , B . H ealy, N . R atliff, G . E ngelm ann, F. L oop, J. R . K ram er, C . K ittrell, and M . S . F eld , P ro c. S P I E ±I nt. S oc. O pt. E ng . 9 06 , 3 20 ( 1988) . 26. K . T. S chom acker, J. K . F r iso li, C. C . C om pton, T. J. F lotte, J. M . R ichter, N . S . N ish iok a, and T. F. D eutsch , L ase rs S u rg er y M ed. 12 , 63 (1 9 92). 27. A . J. Welch, C . G ar dner, R . Richard s-K or tum , G . C r isw ell, E . C han , J. P fefer, and S . Warr en, L ase rs S ur ger y M ed. 2 1, 1 66 (1 9 9 7). 28. R . R ichard s- K ortu m , in O ptical-T herm al R esp onse of L a se r I rra diated T issu e, A . J. Welch and M . Va n G em er t, E ds. (P len u m P r ess, N ew Yo rk , 199 4) , C h ap. 21. 29. T. J. F arell, M . S . P atterso n, and B . C . W ilson , M ed . P h ys. 1 9 , 8 79 ( 1992) . 30. L . W an g and S . L . Ja cqu es, M ed. P hy s. 21, 1 081 (1 994 ). 31. L . W an g and S . L . Ja cqu es, A pp l. O p t. 34 , 2362 (1 995 ). 32. S . P. L in, L . Wan g, S . L . Ja cques, and F. K . T ittel, A p p l. O p t. 3 6, 13 6 ( 199 7) . 33. M . G . N ich ols, E . L . H ull, an d T. F oster, A ppl. O p t. 3 6, 93 (1 9 9 7). 34. J. R . M ou rant, J. J. Bigio, J. Bo yer, R. L . C on n , T. Jo hnso n , an d T. S him ada, L ase r s S urg er y M ed. 17, 3 50 ( 1995) . 35. C . J. L ynn, D . G . O elber g, and S . L . Ja cques, B iol. N eonate 6 4, 69 ( 1993) . 36. A . J. D u rk in , D isse r tation , U n iver sity of Texas at A ustin, A u stin ( 1995) . 37. J. W u, M . S . F eld , and R . P. Rava, A ppl. O pt. 3 2, 358 5 (1 99 3 ). 38. R . A . Z angar o, L . S ilveira Jr., R . M anoh ar an , G . Z onios, I. Itzk an , R . R. D asa r i, J. Va n D am , an d M . S . F eld, A ppl. O p t. 3 5, 5 2 11 ( 1997) . 39. J. A . Z uclich, T. S h im ada, T. R. L oree, I. Bigg io, K . S trobl, and S . N ie, L ase rs L ife S ci. 6, 39 (1 99 4). 40. N S F P o lar P r ogram s U V S pectror adiom eter N etw or k 1 9 94 ±1 99 5 O p erations Repor t; N S F U V R adiatio n M onitoring N etw o r k 19 94 to 1 995, Vo lum e 5 .0 D ata S et. A vailable at w w w .b iosp h erical.co m . 41. E . V. Trujillio, D . R. S andiso n , U . U tzinger, N . R am an u jam , M . F. F ollen- M itchell, and R . Richar ds-K or tum , A p pl. S pectrosc. 5 2 , 9 43 ( 1998) . 42. A . J. D urk in, S . Jaium ar, and R . R ichard s-K o rtum , A ppl. S p ectro sc . 47 , 2114 ( 1993 ). 43. S .-H . C heng, M aster’ s R eport, T he U n iv ersi ty of Texas at A u stin, A u stin (1 99 2). 44. A n I ntro duction to S pectro sc op y for B iochem ists. S . B . Br o w n, E d. ( A cadem ic P r ess, L ond on/N ew Yo r k, 1980 ).


311