INTERNATIONAL JOURNAL OF LEPROSY ^ volume 61, Number 3 Printed in the U.S.A. Clinical and Histopathological Correlation in the Classification of Leprosy' Amarjit S. Bhatia, Kiran Katoch, Rangarajan B. Narayanan, Gopal Ramu, Ashok Mukherjee, and Ravinder K. Lavania 2 Leprosy is an infectious disease primarily affecting the skin and the nerves. The cel- lular characteristics in leprosy lesions are related to the immunological status of the patient. Ridley and Jopling proposed a his- tological classification reflecting the im- munological spectrum ( 8 ), and this classifi- cation has been widely accepted by leprologists and histopathologists. Simul- taneously, clinicians also have adopted the Ridley-Jopling nomenclature and, at pres- ent, the clinical diagnosis is being made along the lines of Ridley-Jopling classification even when a histopathologic examination has not been done. Subsequently, a number of stud- ies have attempted to correlate this histo- logical classification with the clinical picture of leprosy ( 2 ' 3 ' 7 ' 9 ' 10 ). Results of these stud- ies have not been uniform, and in some of these studies the sample sizes have been rather small. Research in leprosy demands accurate classification of the types of leprosy and, in this paper, we report a retrospective analysis of our data in order to assess the concordance between clinical and histo- pathologic classifications in leprosy. MATERIALS AND METHODS During the periods 1979-1981 and 1984- 1987, 1553 skin biopsies from leprosy pa- tients "reported as untreated" were done at Central JALMA Institute for Leprosy ' Received for publication on 2 January 1992; ac- cepted for publication in revised form on 15 April 1993. = A. S. Bhatia, M.Stat., Assistant Director; K. Ka- toch, M.D., Assistant Director; R. B. Narayanan, Ph.D., Ex-Research Officer; G. Ramu, M.D., Ex-Deputy Di- rector; A. Mukherjee, M.D., Deputy Director; R. K. Lavania, M.D., Research Officer, Central JALMA In- stitute for Leprosy, Taj Ganj, Agra 282001, India. Dr. Mukherjee's present address is Institute of Pa- thology (ICMR), Ansari Nagar, New Delhi 110029, India. Dr. Lavania's present address is Railway Hospital, Sholapur, Maharashtra, India. (CJIL), Agra, India, and these patients had been clinically examined and classified ( 8 ) before a biopsy was done. These biopsies were reported on by two pathologists— pa- thologist A during 1979-1981 and pathol- ogist B during 1984-1987. Cases biopsied during 1981-1984 are not included in the analysis. The histopathologic diagnosis of the types of leprosy was made according to the well laid down criteria of Ridley and Jopling ( 8 ). Hematoxylin and eosin as well as Fite-Far- aco stained slides were examined for his- topathology. Among the 1553 cases report- ed, there were some instances in which the clinicians had reported a type 1 or type 2 reaction ( 5 ). The clinical charts and histo- pathology reports of these 1553 patients were re-examined and data pertaining to age, sex, clinical classification of the type of leprosy ( 8 ), histopathologic classification ( 8 ) and reactional status ( 4 ' 5 ) were collected and an- alyzed using standard statistical techniques with the help of a computer program (Dix- on, W. J., et al. BMDP Statistical Software. Los Angeles: University of California, 1987). RESULTS Out of the 1553 cases which have been included in the present study, there were 201 cases classified as in-between types (e.g., BT-BB, BB-BL, etc.) either by the clinicians, the pathologists, or both. These were ex- cluded from the analysis. The remaining cases included one case of neuritic leprosy which also was excluded from analysis. Out of the remaining 1351 cases, 79 were di- agnosed by the clinicians as having reac- tions. The results of concordance for these cases are presented. Table 1 gives the overall association be- tween the clinical and histopathologic di- agnoses of types of leprosy for the 1272 cases which did not include any reactional case or any case with an in-between type of clas- sification. 433
Transcript
INTERNATIONAL JOURNAL OF LEPROSY^ volume 61, Number 3
Printed in the U.S.A.
Clinical and Histopathological Correlation in theClassification of Leprosy'
Amarjit S. Bhatia, Kiran Katoch, Rangarajan B. Narayanan,Gopal Ramu, Ashok Mukherjee, and Ravinder K. Lavania 2
Leprosy is an infectious disease primarilyaffecting the skin and the nerves. The cel-lular characteristics in leprosy lesions arerelated to the immunological status of thepatient. Ridley and Jopling proposed a his-tological classification reflecting the im-munological spectrum ( 8 ), and this classifi-cation has been widely accepted byleprologists and histopathologists. Simul-taneously, clinicians also have adopted theRidley-Jopling nomenclature and, at pres-ent, the clinical diagnosis is being made alongthe lines of Ridley-Jopling classification evenwhen a histopathologic examination has notbeen done. Subsequently, a number of stud-ies have attempted to correlate this histo-logical classification with the clinical pictureof leprosy ( 2 ' 3 ' 7 ' 9 ' 10). Results of these stud-ies have not been uniform, and in some ofthese studies the sample sizes have beenrather small. Research in leprosy demandsaccurate classification of the types of leprosyand, in this paper, we report a retrospectiveanalysis of our data in order to assess theconcordance between clinical and histo-pathologic classifications in leprosy.
MATERIALS AND METHODSDuring the periods 1979-1981 and 1984-
1987, 1553 skin biopsies from leprosy pa-tients "reported as untreated" were done atCentral JALMA Institute for Leprosy
' Received for publication on 2 January 1992; ac-cepted for publication in revised form on 15 April1993.
= A. S. Bhatia, M.Stat., Assistant Director; K. Ka-toch, M.D., Assistant Director; R. B. Narayanan, Ph.D.,Ex-Research Officer; G. Ramu, M.D., Ex-Deputy Di-rector; A. Mukherjee, M.D., Deputy Director; R. K.Lavania, M.D., Research Officer, Central JALMA In-stitute for Leprosy, Taj Ganj, Agra 282001, India.
Dr. Mukherjee's present address is Institute of Pa-thology (ICMR), Ansari Nagar, New Delhi 110029,India.
Dr. Lavania's present address is Railway Hospital,Sholapur, Maharashtra, India.
(CJIL), Agra, India, and these patients hadbeen clinically examined and classified ( 8 )before a biopsy was done. These biopsieswere reported on by two pathologists— pa-thologist A during 1979-1981 and pathol-ogist B during 1984-1987. Cases biopsiedduring 1981-1984 are not included in theanalysis.
The histopathologic diagnosis of the typesof leprosy was made according to the welllaid down criteria of Ridley and Jopling ( 8 ).Hematoxylin and eosin as well as Fite-Far-aco stained slides were examined for his-topathology. Among the 1553 cases report-ed, there were some instances in which theclinicians had reported a type 1 or type 2reaction ( 5 ). The clinical charts and histo-pathology reports of these 1553 patients werere-examined and data pertaining to age, sex,clinical classification of the type of leprosy(8 ), histopathologic classification ( 8 ) andreactional status ( 4 ' 5 ) were collected and an-alyzed using standard statistical techniqueswith the help of a computer program (Dix-on, W. J., et al. BMDP Statistical Software.Los Angeles: University of California, 1987).
RESULTSOut of the 1553 cases which have been
included in the present study, there were201 cases classified as in-between types (e.g.,BT-BB, BB-BL, etc.) either by the clinicians,the pathologists, or both. These were ex-cluded from the analysis. The remainingcases included one case of neuritic leprosywhich also was excluded from analysis. Outof the remaining 1351 cases, 79 were di-agnosed by the clinicians as having reac-tions. The results of concordance for thesecases are presented.
Table 1 gives the overall association be-tween the clinical and histopathologic di-agnoses of types of leprosy for the 1272 caseswhich did not include any reactional caseor any case with an in-between type of clas-sification.
433
434^ International Journal of Leprosy^ 1993
TABLE 1. Comparison between classifications of clinician and pathologist (A and Bconibined).
It can be seen that there was an overallconcordance of 69%. Concordance figureswere very high for LL cases (91%) followedby BT cases (77%). For TT and BL cases,the concordance figures were rather low, 50%and 43%, respectively. Thirty percent of theTT cases were classified as BT and 36% ofthe BL cases were classified as LL by thepathologists. The observed concordance wasminimal for the indeterminate or I (36%)and BB (26%) types of leprosy. The overallconcordance was 75% and 72% when BT/BB, BB/BL (clinical diagnosis) cases wereconsidered as single entities. However, if weconsider TT and BT together, and also BLand LL together, the overall figure of con-cordance was 76%; concordance for the TT/BT group was 80%, for the BL/LL group,93%.
The observed concordance between thehistopathologic and clinical diagnoses wasalso analyzed in relation to the age and sexof the patients. The present data did notshow any evidence of a relationship of these
variables with the concordance between theclinical and histopathologic diagnoses.
Since the histopathologic diagnosis wasmade by two different persons, it may benecessary to consider them individually.Tables 2 and 3 give the correlation betweenthe clinical and histopathologic diagnoseswith respect to pathologists A and B. Pa-thologist A (Table 2) had classified 884 bi-opsies, the remaining 388 were classified bypathologist B (Table 3).
Pathologist A versus clinician. Out of the884 cases classified by pathologist A, no ev-idence of leprosy was seen in 46 (5%) bi-opsies. The overall concordance rate was63%. A maximum amount of concordancewas observed for LL (86%) and BT (75%).A poor rate of concordance was seen in BB(26%) and I (34%) cases; 23 out of 61 (38%)of the I cases were classified as "no leprosy"by the histopathologist. Combining TT withBT cases and BL with LL cases, the overallconcordance rate was 72%. For the TT/BTand BL/LL groups, the agreement between
TABLE 2. Comparison between classifications of clinician and pathologist A.
the clinician and pathologist A was ob-served in 80% and 91%, respectively.
Pathologist B versus clinician. Patholo-gist B classified 388 biopsies; in 22 (6%), noevidence of leprosy was reported. An over-all concordance of 84% was observed be-tween pathologist B and the clinician. A highorder of concordance was observed for LLcases (98%) followed by 82% for BT cases.Minimal concordance was observed for theI and BB types of leprosy; for 61% of theindeterminate cases, pathologist B did notfind any evidence of leprosy. The combiningof TT with BT cases and BL with LL casesresulted in an overall concordance rate of87%; for the TT/BT group agreement wasobserved in 80%, for the BL/LL group, 98%.
Diagnosis of classification for cases withreactions. Since the proportion of cases re-ported as having reactions was low, the con-cordance between the clinical and histo-pathologic diagnoses of classification andreactions was studied, taking the pooledsamples of pathologists A and B. Concor-
dance between the clinical and histopath-ologic diagnoses of the classification of 79cases reported to be "in reaction" by theclinician is given in Table 4. A high levelof concordance was observed for LL andBT cases. For the TT and BL types, sincevery few cases were available, the resultsmay not be meaningful. A fairly good con-cordance was found for borderline reactioncases; overall agreement was observed in81% of the cases.
DISCUSSIONA total of 1272 patients was analyzed to
assess the degree of correlation betweenclinical and histopathologic classificationsof leprosy. The patients were graded clini-cally by at least one of the physicians of theInstitute and histopathologically classifiedby two pathologists. Although the overallpercentages of agreement are somewhat dif-ferent, the pattern of correlation for the var-ious types of classifications is the same forthe two histopathologists as for the clini-
TABLE 4. Comparison between classifications of clinician and pathologists A and Bcombined for cases in reaction.
cian's diagnosis. Although there was no for-mal blinding, the independent nature of thehistopathologic diagnosis is reflected by thesignificant discordance in clinical and his-topathologic diagnoses, especially in bor-derline cases. LL cases seem to present theleast problem for classification. For BL cases,as compared with LL cases, the amount ofagreement was low. A majority of the dis-cordant cases was classified as LL by his-topathology. At the other end of the leprosyspectrum, although fairly good concordancewas observed for the BT type, a low degreeof agreement was seen for TT cases, with afairly good majority of discordant cases be-ing classified as BT by histopathology. Forboth pathologists, minimal agreement wasseen in the I and BB cases.
Our study shows that there are problemsof discordance in the clinical and histo-pathologic diagnoses of some types of lep-rosy (TT, BB and I). Since the same typesof trends were observed among different pa-thologists and clinicians, the observed dis-parity does not appear to be because of thesubjective nature of the observations, eitherclinical or histopathologic. This suggests thatthe criteria of giving weight to certain clin-ical signs versus clinical types or histologicparameters versus clinical types or consid-ering any of these as a "Gold Standard" maynot be ideal.
A sizable proportion of leprosy cases isin a continuously changing immunologicalspectrum, i.e., BT-BB-BL. In some earlycases, clinical signs and symptoms may pre-cede the presently known characteristic tis-sue changes, or vice versa. If a biopsy istaken at an early stage, there is likely to bediscordance between the clinical and his-topathologic observations. Although thesecases were reported as "untreated," thechances that some of these cases had takentreatment earlier cannot be ruled out. Theresults may be affected partially by the stageof treatment. It would be of interest to fol-low proven untreated borderline cases afterchemotherapy.
Tuberculoid leprosy (TT) is only slightlydifferent from BT leprosy, both clinicallyand histopathologically. Clinically, thesecases present as well-defined lesion(s) withpartial or complete loss of sensation with orwithout a thickened nerve and very few acid-fast bacilli (AFB), if any. Histologically, both
of these leprosy types have very well-de-fined granuloma with lymphocytic infiltra-tion and very few AFB. Immunologically,both also are similar to a large extent anddiffer only in degree. Therefore, the line ofdemarcation between TT and BT oftenoverlaps both clinically and histologically.Since many cases diagnosed as TT have his-topathologic evidence of BT, the presentcriteria of differentiating them does not ap-pear to be adequate. It is also possible thatthat proportion of cases belonging to thestable polar tuberculoid type may be muchlower than thought earlier. Therapeutically,TT and BT cases arc treated similarly andso are LL and BL cases. Therefore, com-bining these two groups (TT, BT and LL,BL) does not affect the chemotherapy of thedisease. BB cases present special problemsand even combining these cases with BT orBL does not seem to reduce the discordancesignificantly. Whether they should continueto be considered along with lepromatous(BL, LL) cases, or some changes are nec-essary, cannot be predicted from our presentstudy.
In the work published to date, the overallagreement figures reported a range from alow of 33% ( 9 ) to a high of 77% ( 2 ). It shouldbe noted that the percentage of overallagreement observed in a particular situa-tion, besides depending upon other factors,is a function of the proportion of cases ofclassification of each type of leprosy in thetotal sample. Since the percentage of casesof each type of leprosy is different in variousreported studies ( 3 ' 9 ' 10), the comparisonsmade on the basis of overall percentage ofconcordance may not be valid. Although theoverall agreement observed is also depen-dent upon the individual histopathologistand clinician, a better way of making sucha comparison between various studies is towork out these percentages of agreementseparately for each type of leprosy. For thepurpose of comparison of overall percent-ages between studies, some standardizedprocedure should be adopted.
The indeterminate cases appear to beproblematic due to the nonspecific histologyof their lesions. Definitive indeterminate di-agnosis presently depends upon the dem-onstration of nerve lesion(s) (6 ). PathologistA could not find any evidence of leprosy in38% of indeterminate cases; pathologist B
61, 3^Bhatia, et al.: Clinical-Histopathological Correlation^437
could not demonstrate bacilli in 61% of thesecases. As expected, the ultimate failure/use-fulness of histopathology in suspected andindeterminate leprosy will depend uponseveral factors such as the nature and depthof the biopsy, the quality of the section, andthe number of acid-fast-stained sections ex-amined, etc. Since the Institute is a referralcenter, these aspects are usually kept inmind. Because this was not a formal pro-spective study, some possible influence ofany of these factors cannot be ruled out. Byemploying fluorescence, biochemical andmolecular markers/gene amplification tech-niques, it may be possible to characterizethese cases in a better way in the future. Itmay also be important to study the inter-observer variations in the clinical diagnosesin borderline and indeterminate leprosyversus the results of histopathologic andother markers.
In the reactional cases, the degree of con-cordance between the clinician and the his-topathologist is, by and large, reasonable inthe diagnosis of types. For TT and BT onlya few cases were available. Since the polarvariety of tuberculoid (TT) leprosy is be-lieved to be immunologically stable, thesecases could have been of a subpolar varietyalso. It is, however, a question of debateand subclassification since reactions in tu-berculoid (TT) have been described by someauthors ( 3 ). The observed disagreement,though in a small proportion of cases, callsfor the reappraisal of criteria for identifyingreactions histopathologically as well as clin-ically.
The study reported has a number of lim-itations. Due to the lack of absolute stan-dards, it was not possible to measure thesensitivity or specificity of the diagnosis/classification made by either the clinicianor the histopathologist. The design of thepresent study did not permit us to work outinter-observer variations between the twopathologists and between any two clini-cians. Also, due to limitations in the studydesign, Cohan's Kappa (') could not be re-ported. The present data, being hospitalgenerated, cannot be regarded as represen-tative of the population of leprosy patientsin general.
To conclude, in-depth studies are re-quired to reassess the criteria, giving weightto different clinical signs and histopatho-
logic parameters, in relation to the diagnosisof the different types of leprosy.
SUMMARYThis study reports our observations on
the correlation between clinical and histo-pathological diagnoses of the classificationof leprosy. The histopathological classifi-cation of leprosy in 1351 cases was done perRidley-Jopling criteria and was comparedwith the clinical diagnoses of the same cases.These 1351 cases included 79 cases diag-nosed clinically as having a "reaction."However, the histopathologists could notdetect any evidence of reaction in 16 of these79 cases (20%).
Of the remaining 1272 cases, 68 (5%) werereported as "no evidence of leprosy" by thehistopathologists; 37 of these 68 were foundto be from the clinically indeterminate typeof leprosy. Histopathological and clinicaldiagnoses of the classification of leprosy co-incided in 69% of the cases. Concordancebetween the clinical and histopathologicaldiagnoses for different types of leprosy was:indeterminate (I) = 36%, tuberculoid (TT)= 50%, borderline tuberculoid (BT) = 77%,borderline (BB) = 26%, borderline lepro-matous (BL) = 43%, and lepromatous (LL)= 91%. When some of the types were com-bined (BT with TT, BL with LL), the overallconcordance figure was 76%; concordancefor the TT/BT group was 80%, for the BL/LL group it was 93%. Since both TT andBT are considered paucibacillary and LL orBL are considered multibacillary for treat-ment purposes, differentiating TT from BTor BL from LL is, perhaps, therapeuticallyirrelevant. However, for classification pur-poses it appears that the weight given todifferent signs and/or histopathological pa-rameters for classifying leprosy cases (es-pecially TT, BB and I) needs to be reas-sessed.
RESUMENEste estudio se refiere a nuestras observaciones sobre
Ia correlaciOn entre los diagnOsticos clinico e histo-patolOgico en relaciOn a Ia clasificaciOn de Ia lepra.Siguiendo los criterion de Ridley Y Jopling se hizo laclasificaciOn histopatolOgica de 1351 casos de lepra yse comparO con Ia clasificaciOn clinica de los mismoscasos. Aunque dentro de estos 1351 casos se incluyeron79 diagnosticados clinicamente como casos "en reac-ciOn," los histopatOlogos no pudicron detectar ningunaevidencia de reacciOn en 16 de estos 79 casos (20%).
438^ International Journal of Leprosy^ 1993
Dc los 1272 casos rcstantes, 68 (5%) se reportaroncomo "sin evidencias de lepra" por los histopatOlogos;
37 de cstos 68 casos fueron clinicamentc diagnosti-cados Como lepra indeterminada. Los diagnOsticus cli-
nico c histopatolOgico coincidieron en el 69% de loscasos. La concordancia entre los diagnOsticos clinico e
histopatolOgico para los di ferentes tipos de lepra thecomo siguc: indctcrminada (I) = 36%, tuberculoide
termedia (BB) = 26%, lepromatosa subpolar (BL) =43%, y lepromatosa (LL) = 91%. Cuando se combi-
naron algunos de los tipos (HT con TT, BL con LL),
la cifra de concordancia general fee del 76%; Ia con-cordancia para el grupo TT/I3T foe del 80%, micntrasque para el grupo I3L/LL fuc del 93%. Pucsto que para
propOsitos de tratamiento tanto TT como BT se con-sidcran paucibacilares y I3L y LL se considcran mul-
tibacilares, diferenciar entre TT y I3T o entre BL y LL
(BL) = 43%, ct lepromateux (LL) = 91%. Quand on acombine certains types (BT avec TT, I3L avec LL), letaux de concordance globale êtait de 76%; la concor-
dance pour Ic groupe TT/I3T etait de 80%, ct pour lcgroupe I3L/LL de 93%. Puisquc TT et BT sont toils
deux consideres comme paucibacillaires et LL et I3Lcomme multibacillaires en cc qui concerns le traite-ment, differencier TT de BT ou BL do LL est pout-etre
sans interet sur Ic plan therapeutique. Cependant, dans
des huts de classification. it apparait que le poids donnea diflirents signes ct/ou parametres histopathologiquespour classer les cas do lepre (partictilierement TT, BB
et I) a besoin d'etre reevalue.
Acknowledgment. The authors are indebted to Dr.
H. Srinivasan, Director, and Dr. V. M. Katoch, As-sistant Director, Central JALMA Institute for Leprosy,Agra, for their valuable comments on a previous draft.
The wordprocessing help provided by Mr. Arun Ku-mar I3abar and the secretarial help of Mr. J. D. Kush-
wah are also acknowledged. The referees of the sub-
mitted manuscript are gratefully acknowledged for their
useful comments.
REFERENCESCOHEN, J. A. A coefficient of agreement for nom-
inal scales. Educ. Psycho. Measure. 20 (1966) 37-
46.
2. FINE, P. E. NI., Jon, C. K., MCDOUGALL, C. M.,
MEYERS, W. H. and PONNIGHAUS, J. M. Com-
parability among histopathologists in the diag-nosis and classification of lesions suspected of lep-
rosy in Malawi. Int. J. Lepr. 54 (1986) 614-625.
3. JERTI!, V. P. and DESAI, S. R. Diversities in clin-
ical and histopathologic classification of leprosy.
Lcpr. India 54 (1982) 130-134.
4. JOB, C. K. and DHARMENDRA. Histopathology ofreacting skin in leprosy. In: Leprosy, I "o/. 2. Dhar-
mendra, ed. Bombay: Sawant & Company, 1985,
pp. 823-827.5. Jopling, W. H. Leprosy reactions (reactional
states). In: handbook of Leprosy. 3rd edn. Trow-
bridge, U.K.: Redwood Burn Limited, 1984, pp.
68-77.
6. Liu, T.-C., Yen, L.-Z., Ye, G.-Y. and DUNG, G.-J.
Histology of indeterminate leprosy. Int. J. Lcpr.
50 (1982) 172-176.
7. Meyers, W. H., Heggie, C. D., Kay, T. L., Staple,
E. M. and KVERNES, S. The Ridley-Jopling five
group classification of leprosy—correlation of pa-rameters of classification in 1429 leprosy patients.Int. J. Lepr. 47 (1979) 683-684.
8. Ridley, D. S. and JOPLING, W. H. Classification
of leprosy according to immunity; a five-groupsystem. Int. J. Lcpr. 34 (1966) 255-273.
9. SEHGAL, V. N., REGE, V. L. and REY, M. Corre-
lation between clinical and histopathologic clas-sification in leprosy. Int. J. Lcpr. 40 (1977) 278-
280.10. SINGH, K., BeENGAg, B. and SINGH, R. Variations
in clinical and histopathologic classification oflep-
rosy: a report and possible explanation. Lcpr. In-dia 52 (1983) 472-479.
