876
`
`7 Clin Pathol 1995;48:876-878
`
`A “quickscore” method for
`immunohistochemical semiquantitation:
`validation for oestrogen receptor in breast
`carcinomas
`
`S Detre, G Saccani Jotti, M Dowsett
`
`Abstract
`increasingly
`Immunohistochemistry is
`used in the assessment of markers for
`breast cancer prognosis. Semiquantitation
`is frequently desirable but, other than by
`the use of image analysis, the approaches
`currently in use are cumbersome. The
`most common method used is the H-score
`which takes into consideration the staining
`intensity in conjunction with the per-
`centage ofcells staining positively in breast
`carcinoma tissue. A “quickscore” has been
`developed which dispenses with the need
`to count individual cells. The quantitative
`biochemical Abbott enzyme immunoassay
`(EIA) and the Dako immunohistochemical
`assay (IHA) incorporating a semiquan-
`titative H-score, have been used as stand-
`ards against which the IHA quickscore for
`the semiquantitation ofoestrogen receptor
`expression was tested. A good correlation
`was found between the quickscore and the
`ELA, which was as good as that between
`the H-score and EIA. The quickscore is a
`valid approach andthere is no advantage
`in using the more rigorous H-score. A
`positive cut off quickscore of >3 has been
`suggested.
`(Ff Glin Pathol 1995;48:876-879)
`
`Immunohistochemistry, quickscore, oes-
`Keywords:
`trogen receptor, breast cancer.
`
`involves the addition of scores for staining in-
`tensity and the proportion of positive cells,
`which is a departure from the H-score which
`is based on multiplication of these components
`and intuitively is more appropriate if staining
`intensity reflects antigen concentration. The
`aim of this study was to assess whether a mo-
`dification of the quickscore to a multiplicative
`form wasvalid and could provide an acceptable
`alternative to the H-score. Two extra categories
`for the proportion of cells staining positively
`were also included to permit a completely neg-
`ative result, and to take into account the pres-
`ence of very small numbers ofpositive cells.
`
`Methods
`Ninety six primary breast cancer surgical speci-
`mens, embeddedin paraffin wax, were studied.
`These tumours were taken from a cohort of
`119 untreated patients studied previously’ and
`the oestrogen receptor EIA, using the Abbott
`H222 antibody, and the oestrogen receptor
`THA,using the Dako 1D5 antibody, have been
`described in detail. The reduction in sample
`size to 96 tumours was necessary because 17
`mismatches had occurred between EIA and
`THA pairs of results and in order not to con-
`found the scoring appraisal these mismatches
`were excluded from this investigation. A further
`six tumours were excluded because full results
`were not available.
`
`Oestrogen receptor determinationis important
`in the study of the biology of breast cancers
`and for determiningthe clinical implications of
`hormone treatment,’ Recently, the immuno-
`histochemical assay (THA) has gained favour
`over the gold standard enzyme immunoassay
`(EIA) becauseit is cheaper, quicker and easier
`to perform and enables direct visualisation of
`malignant tumour receptorlocalisation in small
`tissue samples.* One disadvantage of the im-
`munohistochemical approach is the laborious
`H-score procedure which is prone to a degree
`of discordance between assessors.°
`As the immunohistochemical approach is
`gaining increasing support
`for
`routine as-
`sessment of oestrogen receptor expression, a
`more rapid scoring scheme would be useful
`especially when appraising large numbers of
`sections from clinical studies. To this end a
`quickscore method such as that suggested by
`Barnes et al* would expedite matters. The
`method described by Barnes er al, however,
`
`SCORING ASSESSMENT
`
`For H-score assessment, 10 fields were chosen
`atrandom at x 400 magnification andthestain-
`ing intensity in the malignant cell nuclei was
`scored as 0, 1, 2, or 3 corresponding to the
`presence of negative, weak, intermediate, and
`strong brown staining, respectively. The total
`number ofcells in each field and the number
`of cells stained at each intensity were counted.
`The average percentage positive was calculated
`and the following formula?’ was applied:
`IHA H-score=(% ofcells stained at intensity category
`1x1)+(% of cells
`stained at
`intensity category
`2™2)+(% of cells stained at intensity category 3 x 3).
`
`A H-score berween 0 and 300 was obtained
`where 300 was equal to 100% of tumour cells
`stained strongly (3+).
`The immunohistochemically stained sec-
`tions were subsequently given a quickscore
`independently by two of the authors without
`prior consultation or recourse to clinical, bio-
`
`Genome& Co. v. Univ. of Chicago, PGR2019-00002
`UNIV. CHICAGO EX. 2077
`
`ulham Road,
`London SW3 6)J
`S Detre
`M Dowsett
`
`Istituto di Anatomia
`ed Istologia
`Patologica,
`University of Parma,
`Italy
`G Saccani Jotti
`Correspondence to:
`S Detre.
`Accepted for publication
`17 January 1995
`
`

`

`Quickscore for immunohistochemical semiquantitation
`
`877
`
`s
`A
`++
`Table 1 Spearman rank correlation coefficients for
`scores and a good positive correlation was ob
`SOMDLPIONS
`a hs
`Varios
`tain Wf outhie Macete
`served between the IHA andthe EIA (table 1).
`UeESlireERY ti PAGE Cannaettcrs
`—_——— _Moore importantly,the correlation between both
`Looapeanicios af weattion’s ofsatoricnohene
`rho,
`of the quickscores and the EIA wasvery similar
`
`Q=AxB wvHscore 0-892 to that for the H-score and the EIA.
`
`eeeee
`Coes
`We also compared the performance ofthe
`Q=AxBwEIA
`0-831
`two quickscore approaches with two separate
`Q=A+B ov EIA
`0-832
`eeeCCsaasssessors. Total agreement of score occurred
`rho,=regression coefficient corrected for ties; Q=quickscore
`in 73% of cases and demonstrated a very strong
`WeelDoone aan receptor positive malignant
`correlation between the multiplicative and ad-
`p<0-0001
`ditive quickscore (rho,=0-994). In the seven
`cases where disagreement occurred, five were
`at
`the highest
`score levels—for example,
`chemical or previously recorded H-score data) Ax B=18 and 15, A+B=8 and9, and two at
`The quickscore categories were based on both
`the lowest levels—for example, Ax B=1 and
`the intensity and the proportion ofbrownstain-
`2, A+B=2 and 3. The maximum quickscore
`ing nuclei. In our modification the proportion was attained in many cases, especially where
`of malignantcells staining positively through-
`high EJA values occurred. Interestingly, a very
`out the section was termed category Aand was
`high H-score level was never achieved.
`assigned scores from 1 to 6 (l1=0-4%; 2=
`5-19%; 3 = 20-39%; 4=40-59%; 5 = 60-79%;
`6=80-100%). The whole section was scanned Discussion
`at low powerin order to gauge the general level The concordance between the two observers
`of intensity throughout. The average intensity,
`indicates that both quickscores are repro-
`corresponding to the presence of negative,
`ducible although more studies from other
`weak, intermediate, and strong staining, was
`laboratories are necessary to confirm this. The
`given a score from 0 to 3, respectively, and
`THA quickscore compared favourably with the
`termed category B. Category A was added to
`gold standard EIA. That the quickscore is of
`category B (A+B) to form an additive quick-
`equivalent value to the H-score, when using
`score and recorded in parallel with the product
`the same immunohistochemical technique, was
`(A xB) as a multiplicative quickscore. The res-
`confirmed by the good Spearman rank cor-
`ults were tabulated and EIA and THA results_relation between them.
`were compared with the quickscore using
`The reason why these sections did not have
`Spearman rank correlation statistics.
`very high H-scores could have been because
`tumours in this cohort expressed oestrogen
`receptor at a low level. This, however, was not
`the case when the quickscore was applied as
`Results
`The EIA values ranged from 0 to 597 fmoles/ maximum scores were reported in 19% (GSJ)
`mg protein and 62% of the tumours were
`and 14% (SD)ofcases.
`oestrogen receptorpositive (cut off > 10 fmoles/
`The quickscore seems to be less prone to
`mg protein). H-scores ranged from 0 to
`variation than the H-score becausethe observer
`190, Additive and multiplicative quickscores
`evaluates the whole section, estimating an over-
`covered the full range—that is, (A+B)=1-9
`all impression of intensity when scoring and not
`and (A x B) =0-18, respectively. Table 1 shows
`just 10 representative, but randomly selected,
`acomparison ofthe Spearmanrankcorrelations
`high power fields. This difference in approach
`for the three oestrogen receptor methods and might explain the underestimation of H-score.
`the individual categories of the AxB quick-
`The positive cut off for the EIA is usually
`score are compared with intervals of H-scores
`10 fmoles/mg protein in breast carcinomas.**’
`and EIA in table 2. There was no difference The positive cut off using the IHA is more
`between the additive and multiplicative quick-
`contentious and varies from a H-score of 20,
`
`
`
`0
`2
`0
`0
`0
`0
`
`
`
`0
`0
`2
`0
`0
`0
`
`
`
`0
`0
`1
`3
`0
`0
`
`
`
`9
`0
`3
`5
`0
`0
`
`29
`2
`0
`0
`0
`0
`
`0
`3
`Sed
`0
`0
`0
`
`Table 2 Frequency of multiplicative quickscores in JHA and ELA result subsets
`Quickscom=A x B
` 0 I 2 # He & 10 12 5 18 Total
`
`H-score
`0
`1-19
`20-19
`50-99
`100-199
`200-300
`EIA
`36
`0
`0
`0
`0
`0
`0
`0
`2
`3
`31
`so
`24
`2
`2
`4
`5
`a
`2
`1
`i 0
`0
`10-49
`16
`Mt
`|
`4
`5
`1
`2
`i
`0
`0
`0
`50-99
`12
`5
`1
`5
`I
`0
`0
`0
`0
`0
`0
`100-199
`5
`3
`0
`1
`1
`0
`0
`0
`0
`0
`0
`200-299
`1
`1
`0
`i)
`0
`0
`0
`0
`0
`0
`0
`300-399
`0
`0
`0
`0
`9
`0
`0
`o
`0
`0
`0
`400-199
`2
`1
`0
`Oo
`1
`0
`0
`0
`9
`0
`0
`500-600
`
`
`
`
`
`
`
`
`
`
`31 4 2 2. 4 8 13 14 4 14Total 96
`
`*=32; **=29 fmoles/mg protein.
`.
`Negative H-score < 19.
`Negative ELA <9 fmoles/mg protein.
`
`
`
`0
`0
`0
`5
`8
`Q
`
`
`
`0
`0
`0
`4
`10
`0
`
`
`
`0
`0
`0
`1
`3
`0
`
`
`
`0
`0
`0
`2
`12
`0
`
`
`
`29
`7
`7
`20
`33
`0
`
`

`

`in our laboratory,’ to 50-100in others.”’ Bear-
`ing in mind our laboratory cutoff values, table
`2 shows a logical demarcation between the
`positive and negative EIA and THA subsets
`at a multiplicative quickscore >3. Thirty six
`breast
`tumours with negative EIA results
`(<9 fmoles/mg protein) occurred in the 0-2
`quickscore subset; no oestrogen receptor neg-
`ative breast tumours on EIA were reported in
`the 3-18 quickscore subset whereas 59 of 60
`oestrogen receptor positive breast cancers were
`observed in this classification. Only one of 60
`oestrogen receptor EIA positive tumours was
`presentin the 0-2 quickscore subset, The same
`results were obtained using a H-score cut off
`<20. The one tumour negative on assessment
`using the quickscore but positive by IHA (H-
`score=32)
`and
`oestrogen’
`receptor-EIA
`(29 fmoles/mg protein) came from the same
`patient,
`Theresults also show that there is no differ-
`ence between adding or multiplying categories
`Aand B. The additive approach, with a range
`from 1 to 9, has two disadvantages. Firstly, the
`additive (A+B) quickscore does not permit a
`completely negative result, which occurred in
`many of our oestrogen receptor negative cases,
`whereas the multiplicative (Ax B) quickscore
`does. Secondly, the range is small for the ad-
`ditive quickscore. Considering the wide range
`
`Detre, Saccani Forti, Dowsert
`
`of values within the quantitative methods, the
`multiplicative quickscore, ranging from 0 to
`18, is preferable. Without prejudicing accuracy,
`the quickscore approach takes about a quarter
`of the time taken when calculating the H-score
`because it dispenses with the need to count
`individual cells.
`
`G Saccani Jotti was supported in part by the following grants:
`CNR ACRO Project, CNR Bilaceral 94.02482.CT04 and
`MURST 40% (Italy). SD and MD were supported by the
`Cancer Research Campaign,
`
`1 Osborme CK. Receptors. In: Harris JR, Helman $, Henderson
`IC, Kinne DW, eds. Breasr diseases. Philadelphia: JB Lip-
`pincott, 1991:301-25.
`2 Andersen J. Determination of oestrogen receptors in paraffin-
`embedded tissue. Acta Oncol 1992;31:611-27.
`3 McClelland RA, Finlay P, Walker KI, Nicholson D, Rob-
`ertson JFR, Blamey RW, er al. Automated quantitation
`of immun
`cally localised oestrogen receptors in
`human breast cancer. Cancer Res 1990;50:3545-50,
`4 Barnes DM, Dublin EA, Fisher CJ, Levison DA, Millis
`RR. Immunochistochemical detection of p53 protein in
`mammary carcinoma. Him Pathol 1993;24:469-76.
`5 Saccani Jormi G, Johnston SRD, Salter J, Detre 5, Dowsert
`M. Companson of a new immunohistochemical assay for
`oestrogenreceptor in paraffin wax embedded breast car-
`cinoma tissue with quantitative enzyme immunoassay. 7
`Glin Pathol 1994;47:900-5.
`6 Andersen J, Bentzen SM,Poulsen HS. Relationship between
`radioligand binding
`assay, immunoenzyme assay and im-
`munohistochemical
`assay for oestrogen receptors in human
`breast cancer and association with rumour differentiation.
`Eur J Cancer 1988;24:377-84.
`7 Kinsel B, Szabo E, Greene GL,Lrgeine J, Leight GS,
`McCarty KS. Immunocytochemical analysis of oestrogen
`receptors as a predictor of prognosis in breast cancer
`patients:
`co
`on with
`quantitative
`biochemical
`methods. Cancer Res 1989;49:1052-6,
`
`9 Clin Pathol 1995;48:878-880
`
`Granulomatous bone marrow inflammation
`during treatment of chronic myeloid leukaemia
`with interferon alpha-2b
`
`A Siboni, T Mourits-Andersen, J Moesner
`
`Abstract
`study, partial cytogenetic response and tem-
`porary morphological remission was seen in
`A patient with chronic myeloid leukaemia
`46/71 patients.' Interferon alpha is used in the
`developed bone marrow granulomas dur-
`treatment of both infectious and malignant
`ing treatment with interferon alpha-2b.
`diseases.” One of its mechanismsof action may
`Some granulomas had necrotic centres
`beinhibition offormation ofnew bloodvessels
`and giant cells and there was marked eos-
` (antiangiogenesis) in tumour areas.”
`inophilia surrounding them. The gram-
`Bone marrow necrosis at blast
`trans-
`ulomas disappeared when the interferon
`formation of chronic granulocytic leukaemia
`treatment was
`discontinued. Myco-
`treated with interferon has been described by
`bacteriosis was ruled out. The most likely
`explanation for the granuloma formation Kendra eral,’ but the formation ofbone marrow
`was drug hypersensitivity.
`granulomas during the treatment of chronic
`(J Clin Pathol 1995;48:870-880)
`myeloid leukaemia with IFN-«-2b has not been
`described before.
`
`Keywords: interferon, bone marrow granulomas.
`
`The treatment of chronic myeloid leukaemia Case report
`with interferon alpha-2b (IFN-a-2b) prolongs
`A 43 year old man was referred in August 1992
`because ofa leucocyte count of89 x 10°/1. Bone
`survival if cytogenic response occurs: in one
`
`Hospital, ome?
`DK-6700 Esbjerg,
`ateteeticine t
`and Haematology
`A Siboni
`T Mourits-Anderson
`Department of
`Pathology
`] Moesner
`Correspondence 10:
`Dr Anders Siboni,
`Skolebakken 66,4,th.
`ieSO Bee
`Accepted for
`publicati
`1 December 1994
`Dew {994
`
`