21/07/2017

British Division of the International Academy of Pathology (BDIAP) KRISTIN HENRY LECTURE, BELFAST 2017: TYPING OF OVARIAN AND ENDOMETRIAL CARCINOMAS: A STORY OF SUCCESS AND FAILURE

W Glenn McCluggage Belfast, United Kingdom

The BDIAP Kristin Henry Lecture Presidents of the IAP from the UK 1982-90 Treasurer of BDIAP 1995-96 President of BDIAP 1997-98 Cunningham Medal 2009-10 President’s Medal 2005-08 Chair Finance Committee 2008-11 Chair Education Committee

Roger Cotton Kristin Henry 1982-1983 2010-2012

TYPING OF OVARIAN AND ENDOMETRIAL CARCINOMAS: A STORY OF SUCCESS AND OVARIAN CARCINOMA TYPING FAILURE • Ovarian carcinoma- typing • Recent developments regarding “ovarian” •SUCCESS serous carcinomas • Endometrial carcinoma- typing • Small cell carcinoma of the of hypercalcaemic type (SCCOHT)

1 21/07/2017

REPRODUCIBILITY OF OVARIAN OVARIAN EPITHELIAL CARCINOMA CARCINOMA TYPING (WHO 2014)- 5 major histotypes • Brugghe et al (IJGC, 1995) - 61% • low grade serous • Bertelsen et al (IJGC, 1993) - 72% (serous and • high grade serous endometrioid); 86% (mucinous); 100% (clear cell) • mucinous • Baak et al (AQCH, 1986) - significant variation • endometrioid • Cramer et al (APLM, 1987) – suboptimal • Lund et al (APMIS, 1991) – 68% • clear cell • Sakamoto et al (Gynecol Oncol, 1994) -53% • Brenner • ICON 5- POOR ++ • seromucinous • undifferentiated

DIFFERENCES FROM PREVIOUS STUDIES RECENT POPULATION BASED (WASHINGTON / BRITISH COLUMBIA) • increase in serous • • 68-71% serous (17-18:1 ratio HGSC to LGSC) decrease in endometrioid • • 3% mucinous marked decrease in mucinous (exclusion of secondaries) • 9-11% endometrioid • reversal of endometrioid and clear cell ratio • 12-13% clear cell 88% OF ADVANCED STAGE OVARIAN CARCINOMAS ARE SEROUS

REASONS FOR INCREASE IN SEROUS IMPORTANCE OF TYPING OF OVARIAN AND DECREASE IN ENDOMETRIOID CARCINOMA • Recognised that many that were • at present typing of some (but increasing) previously diagnosed as high grade therapeutic significance; treatment traditionally endometrioid are actually high grade serous more dependent on stage and grade carcinoma (HGSC) • NOW DIFFERENT THERAPY FOR DIFFERENT TYPES (surgery, traditional chemotherapy, • Importance of WT1- Al-Hussaini et al. targetted therapies, hormone therapy) Histopathology 2004; 44; 109-115. • carcinoma grading (differs between different types) • screening, personal and family history risk (BRCA, Lynch)

2 21/07/2017

OVARIAN CARCINOMA GRADING- INTERNATIONAL COLLABORATION ON INTERNATIONAL COLLABORATION ON CANCER REPORTING (ICCR) CANCER REPORTING-ICCR • serous- low grade and high grade (2 different • Alliance between RCPath UK, RCPath Aus, CAP, types) ESP, Canadian Partnership Against Cancer • endometrioid- like uterine corpus • Purpose of developing standardised evidence • clear cell- automatically high grade based cancer datasets for each site • mucinous- as for endometrioid • Ovarian/ fallopian tube/ primary peritoneal dataset- Modern Pathology 2015;28;101-122 • http://www.iccr-cancer.org/datasets

14

RECENT STUDY ON REPRODUCIBILITY OF HISTORIC AREAS OF DIAGNOSTIC OVARIAN CARCINOMA TYPING DIFFICULTY • Koebel et al, Am J Surg Pathol 2010;34;984-93. • Serous (high grade) versus endometrioid • excellent agreement • Serous (high grade) versus clear cell • participants had training in modern criteria • Serous (high grade) versus transitional • important for subtype specific ovarian cancer • Serous (high grade) versus undifferentiated treatments • CONTRAST WITH UTERINE CARCINOMA • good marker- WT1

HIGH GRADE SEROUS CARCINOMA- MORPHOLOGIC DIVERSITY • Papillary/ micropapillary • Slit-like • Glandular • Microglandular/ microcystic • Cystic • Solid • Multinucleate cells • Signet ring cells • Oncocytic/ clear cell • OFTEN ADMIXTURE OF PATTERNS

3 21/07/2017

PSEUDOENDOMETRIOID MORPHOLOGY IN HGSC CLEAR CELL CHANGE IN HGSC (SOMETIMES POSTCHEMOTHERAPY)

TRANSITIONAL-LIKE MORPHOLOGY IN OVARIAN HIGH GRADE SEROUS HGSC CARCINOMA

WT1 in PSEUDOENDOMETRIOID AREAS IN HGSC WT1 in CLEAR CELL AREAS IN HGSC

p53 WT1

4 21/07/2017

WT1 IN TRANSITIONAL-LIKE AREAS IN UNDIFFERENTIATED OVARIAN HGSC CARCINOMA

WT1

MIXED OVARIAN CARCINOMAS Mixed Ovarian Carcinomas

• historically quite common (up to 10-20%) • Improved recognition of types has virtually abolished • most commonly historically reported were mixed mixed tumours – 15 of 871 cases reviewed (1.7%) by H/E using modern serous/ endometrioid; mixed serous/clear cell; diagnostic criteria mixed serous/ transitional; mixed • 22 cases thought to be mixed were investigated further by serous/undifferentiated (mostly variants of high immunohistochemistry and molecular testing grade serous carcinoma- doubtful if these • Only 13 true mixed carcinomas when combinations exist) immunohistochemistry and molecular data incorporated • occasionally get mixed endometrioid/ clear cell • Mixed carcinomas account for less than 1% of ovarian (association with endometriosis) carcinomas MacKenzie et al. Am J Surg Pathol 2015; 39: 1548-1557 • occasionally others • CATEGORY DROPPED FROM WHO 2014

MORPHOLOGY OF BRCA ASSOCIATED TRANSITIONAL CARCINOMA HIGH GRADE SEROUS CARCINOMAS • dropped from WHO 2014 • SET pattern (Solid, pseudoEndometrioid, • most are variants of high grade serous Transitional) (“transitional-like”) (more typical areas if sample • Higher mitoses; more geographical necrosis well; omental metastases may look like high • ? Increased tumour intraepithelial T lymphocytes grade serous; WT1 positive; aberrant p53) • Different patterns of metastatic disease • some are variants of endometrioid carcinoma (“pushing” rather than “infiltrative”) (transitional-like morphology) • Similarities to breast carcinomas in BRCA • • still categories of benign, borderline and PROBABLY ALL PATIENTS WITH HGSC SHOULD UNDERGO BRCA TESTING- FUNDING QUESTIONS malignant Brenner

5 21/07/2017

HGSC WITH CLEAR CELL CHANGE HGSC WITH CLEAR CELL CHANGE VERSUS CLEAR CELL CARCINOMA • Look for areas of typical HGSC or clear cell carcinoma • Look for endometriosis • Status of tubes

HGSC WITH CLEAR CELL CHANGE IMMUNOHISTOCHEMISTRY- HGSC VERSUS CLEAR CELL CARCINOMA WITH CLEAR CELL CHANGE • IMMUNOHISTOCHEMISTRY USEFUL • HGSC: mutation-type p53 (95%), block-type p16 (70%), ER (70%- often diffuse), Napsin A negative • Clear cell carcinoma: wild-type p53, patchy p16, ER negative (sometimes focally positive), Napsin A usually positive • HNF1-beta less useful p53 WT1

IMMUNOHISTOCHEMISTRY- CLEAR HGSC VERSUS ENDOMETRIOID CELL CARCINOMA • Many previously diagnosed as high grade endometrioid adenocarcinoma are HGSC • Squamous elements, endometriosis, obvious low grade endometrioid areas, normal tube are suggestive of endometrioid carcinoma

HNF1 beta Napsin A

6 21/07/2017

IMMUNO- HGSC VERSUS HIGH GRADE PITFALL ENDOMETRIOID ADENOCARCINOMA • WT1- HGSC • WT1 (only about 95% of tubo-ovarian HGSCs • p53- wild-type suggestive of endometrioid; positive) some high grade endometrioid carcinomas • High grade ovarian endometroid have mutation-type (may be clonal pattern rarely WT1 positive suggestive acquisition of TP53 mutation • Low grade ovarian endometroid during progression) adenocarcinomas may be WT1 positive

Interpretation of p53 immunohistochemistry p53 No TP53 mutation

• p53 immunohistochemistry- lot of confusion • only consider positive/significant if diffuse strong nuclear immunoreactivity (75-80% cells suggested- associated with Normal missence mutation) • p53 null consistent with serous carcinoma (different type of mutation (nonsense) or deletion resulting in truncated protein Wild type pattern which is not detected by immunohistochemistry) • third pattern of mutation-type staining- cytoplasmic (rare pattern) Nonsynonymous Stopgain Stopgain • most normal tissues and tumours exhibit focal, weak, heterogenous =missense Indel Indel staining (“wild-type” staining) (usually <50%) (about 5% of HGSCs Splicing Splicing with TP53 mutation exhibit wild-type staining) • DON’T REPORT AS POSITIVE OR NEGATIVE- REPORT AS “WILD- TYPE” or “MUTATION-TYPE/ ABERRANT”

p53 overexpression p53 complete absence p53 cytoplasmic

Abnormal; mutation-type J Pathol Clin Res 2016;2:247

“OVARIAN” SEROUS CARCINOMA (OSC) – RECENT DEVELOPMENTS

• two distinct tumour types (called low grade and high grade OSC) (recognised in WHO 2014) • not two grades of same • different neoplasms with different underlying pathogenesis, molecular events, behaviour, prognosis • high grade much more common than low grade (approximately 17-18:1) • use instead of traditional grading schemes

2016: www.thebagp/resources

7 21/07/2017

PATHOGENESIS (LOW GRADE SEROUS) MOLECULAR (LOW GRADE SEROUS)

• low grade arise from pre-existing benign and • Mutations in genes associated with MAPK borderline tumour (probably not all cases) pathway (KRAS or BRAF mutations in approximately half (sometimes early in • micropapillary variant of serous borderline evolution – mutations found in benign and may be intermediate stage in development of borderline tumours; identical mutations in low grade serous carcinoma borderline and malignant areas in same • well-defined adenoma-carcinoma sequence tumour) • KRAS and BRAF mutations are usually mutually exclusive; equivalent effect on tumorigenesis • no TP53 mutations/abnormalities

PATHOGENESIS (HIGH GRADE SEROUS) MOLECULAR (HIGH GRADE SEROUS) • traditionally thought to arise directly from ovarian surface epithelium or epithelium of cortical inclusion cysts; now clear that most • TP53 mutations (early in evolution-seen in early microscopic cases arise from epithelium of distal fallopian tumours eg BRCA1/2) (p53 mutation found in almost 100% high tube grade serous using stringent methods) (IJGP; PMID 26166714) • precursor lesion is serous tubal intraepithelial • BRCA1/2 abnormalities (germline or somatic mutations or carcinoma (STIC) hypermethylation) • no BRAF and only occasional KRAS mutations • doesn’t arise from borderline tumour

8 21/07/2017

DISTINCTION BETWEEN LOW GRADE AND CLINICAL BEHAVIOUR- LGSC and HGSC HIGH GRADE SEROUS CARCINOMA • LGSC (mean 52) younger than HGSC (mean 62) • Important distinction for management • HGSC poor prognosis (usually presents at advanced purposes stage, responds well initially to chemo but usually recurs; occasional long term survivors) • Usually straightforward • LGSC more indolent (good prognosis for early stage; • Occasionally difficult (especially in small advanced stage- patients usually die of tumour) (IJGP biopsies) 2013; 32; 529-535- stage 2-4, <30% survival at 10 years; not significantly different to HGSC); (AJSP • Only reliable marker is p53 2016;40;627-635; 5 year survival 62.3% LGSC, 43.9% HGSC but no survival difference at 10 years)

HIGH GRADE SEROUS MIMICKING MICROPAPILLARY ARCHITECTURE IN LOW GRADE HGSC

p53

LGSC WITH “BIGGER” NUCLEI LGSC

9 21/07/2017

TRANSFORMATION LOW GRADE INTO HIGH GRADE • rare (AJSP 2012; 36; 368-375) (serous borderline or low grade serous carcinoma) • can be misdiagnosed (bigger nuclei in low grade serous) • can transform to HGSC, anaplastic carcinoma, carcinosarcoma • p53 NOT reliable in such cases (often not p53 mutated)

p53 in LGSC

TRADITIONAL VIEW OF PATHOGENESIS OF HIGH GRADE SEROUS CARCINOMA

• traditionally thought to arise directly from ovarian surface epithelium (OSE) or epithelium of cortical inclusion cysts • thought that some cases might arise from secondary Mullerian system (peritoneum) • doesn’t arise from borderline tumour

IS TUBAL FIMBRIA THE ORIGIN OF EXTRAUTERINE HIGH GRADE SEROUS CARCINOMA?

• proposal that tubal fimbria (distal tube) (secretory cells) is site of origin of many/most extrauterine (ovary, peritoneum and fallopian tube) high grade serous carcinomas

10 21/07/2017

INITIAL EVIDENCE

• came from prophylactic risk reducing salpingo- oophorectomy specimens (RRSO) (BRCA1/2) • once tubes were examined in their entirety, tubal lesions (distal) were occasionally seen with little/ nothing in ovary • tubal lesions may be STIC (serous tubal intraepithelial carcinoma) or small HGSCs • now well established in BRCA patients that tube is origin of HGSCs • ? does same hold true for sporadic HGSCs

p53 in STIC

MIB1 IN STIC p53 SIGNATURES IN TUBE

• small foci of intense p53 immunoreactivity in absence of morphological changes • equally common in BRCA1/2 tubes and in control tubes (occur in all age groups) • most common in fimbria • involves secretory cells • may contain TP53 mutations (? Mutations occurring all the time but most spontaneously regress) • don’t diagnose STIC in isolation in absence of morphological features and confirmatory p53 and MIB1 • overinterpretation of normal fallopian tube morphology; “dangerous” to do p53 on “normal” tubes

11 21/07/2017

WHAT ABOUT SPORADIC HGSCs

• Usually present at advanced stage • Tube (s) often obliterated and embedded in tubo-ovarian mass • Difficult to study precursor lesions • STIC/ mucosal HGSC found in carefully sectioned tubes (when both visible) in significant percentage of cases (up to two- thirds) of sporadic HGSC

p53

Implications for specimen handling

• SEE-FIM protocol ESSENTIAL for identifying STIC/tubal involvement- now routinely done in tubo-ovarian HGSC

BUT IS TUBAL LESION PRIMARY OR ? FIELD-EFFECT IN HGSC METASTATIC? • same TP53 mutations in HGSC at multiple sites • Intramucosal metastasis from a variety of sites may occur in tubes and mimic an in-situ lesion • evidence that clonally related and not part of • Gynaecological or non-gynaecological tumours “field-effect” when spread to tube exhibit mucosal • no/ minimal evidence of field-effect in HGSC involvement and even mimic STIC (AJSP • one site is primary with metastasis to the 2015;39;35-51) others • Some molecular evidence that tube is initial site but difficult to prove by molecular techniques • RECENT OBSERVATIONAL STUDIES PROVIDE FIRM EVIDENCE THAT TUBE IS PRIMARY

12 21/07/2017

GASTRIC TYPE CERVICAL ADENOCA INVOLVING USC INVOLVING FALLOPIAN TUBE TUBE

INCIDENTAL SPORADIC HIGH GRADE SEROUS CARCINOMA • established that incidental tumours in patients with BRCA1/2 mutation are of tubal origin • 4 papers recently published- unsuspected STIC/ HGSC incidentally detected (AJSP 2015; 39; 357- 364) • PROVES that sporadic HGSC of tubal origin (FINAL PIECE OF EVIDENCE)

Am J Surg Pathol, 2014

Summary of findings of incidental HGSC in a non-prophylactic setting

Study Total Cases Invasive Invasive Organ- Organ- Organ- cases with HGSC HGSC in confined confined: confined: STIC in tube ovary Disease tube ovary (tube OR Ovarian? ovary) Peritoneal? Tubal? Undesignated? Rabban, 4 4 3 1 3 3 0 2014

Morrison, 22 22 6 1 21 21 0 …. CHAOS! 2014

Gilks, 2014 21 20 12 2 18 18 0

Total 47 47 22 4 43 43 0

13 21/07/2017

EXTRAUTERINE HIGH GRADE SEROUS SURVEY: INTERNATIONAL JOURNAL OF CARCINOMA- SITE OF ORIGIN GYNECOLOGICAL PATHOLOGY – PMID 27801755 • FIGO 2014- same staging system (ovary, tube, peritoneum, undesignated) • 173 respondents • FIGO 2014 and WHO 2014- no recommendations regarding • Widespread acceptance of tubal origin (86% designating site of origin • WHO- the decision as to primary site should be pragmatic, based on pathologists, 92% clinicians) experience and professional judgement • Clinicians thought it more important to • DOMINANT MASS THEORY TRADITIONALLY USED (ovary designated as primary site in most cases) correctly assign a primary site than • possibilities- pelvic high grade serous; extrauterine; Mullerian; pathologists (71% versus 49%) tubo-ovarian; undesignated • implications:- epidemiology, tumour incidence/mortality, cancer registries, entry into clinical trials • different viewpoints- STIC/ in situ criteria; dominant mass criteria

PROPOSAL FOR DESIGNATING SITE OF ORIGIN Summary of site assignment guidelines proposals OF HGSC Criteria Primary site Comment • extensive examination of tube (SEE-FIM) STIC present Fallopian tube Regardless of presence and • any STIC or mucosal serous ca in tube- tubal origin size of ovarian and peritoneal • if fallopian tube or fimbria not identified (obliterated by mass)- tubal disease origin Invasive mucosal carcinoma Fallopian tube Regardless of presence and • ovarian primary if tumour in ovary and nothing in mucosa of tube (STIC or in tube, with or without STIC size of ovarian and peritoneal invasive) (both tubes need to be clearly visible and examined by SEE-FIM disease protocol) Distal end or entire tube Fallopian tube Regardless of presence and • primary peritoneal- nothing in tube or ovary (vanishingly rare- will likely disappear) (WHO 2014) incorporated into ovarian size of ovarian and peritoneal mass disease • post-chemo (if no residual) or on small biopsy- designate as tubo-ovarian • USING THESE CRITERIA- approximately 80% tubal primaries No STIC or invasive mucosal Ovary Regardless of presence and carcinoma in either tube in size of peritoneal disease • undesignated- very small proportion presence of ovarian mass • (Histopathology 2014; 65; 149-154; Gynecological Oncology 2016;141;195- 198; International Journal of Gynecological Pathology 2016;35;230-237) Both tubes and both Primary peritoneal HGSC As recommended in WHO 57 • CRITERIA ADOPTED BY ICCR grossly and microscopically blue book 2014 normal or involved by benign process in presence of peritoneal HGSC

Basis for tubal assignment in 44 chemonaive RETROSPECTIVE (n=151) PROSPECTIVE (n=111) cases Primary T O P U T O P U site Chemo 63 16 0 1 44 9 0 0 Criterion Number (%) naive (79%) (20%) (0%) (1%) (83%) (17%) (0%) (0%) STIC only 5 (11%) Invasive mucosal +/- STIC 26 (59%) Post- 48 16 7 0 44 7 4 3 Entire tube or part of tube 13 (30%) NACT (68%) (22%) (10%) (0%) (76%) (12%) (7%) (5%) incorporated in mass Total 44

Singh et al, 2015

14 21/07/2017

Unilateral versus bilateral tubal and General implications ovarian involvement in HGSC • Unilateral Bilateral Total Consistency and uniformity in reporting and in message to patients Ovary 18 (38%) 29 (62%) 47 Fallopian 37 (84%) 7 (16%) 44 • Tumour registry data tube • Sharing data and good practice

• Bilaterality = strong indicator of secondary spread • In 53 chemo-naïve cases, ovarian involvement in HGSC was significantly more frequently bilateral • Tubal involvement unilateral in 84% supporting primary rather than metastatic involvement

Singh et al, 2015

ENDOMETRIAL CARCINOMA TYPING ENDOMETRIAL CARCINOMA TYPING •FAILURE •FAILURE (BUT PROMISE OF BETTER TO COME)

DUALISTIC PATHWAY OF ENDOMETRIAL Type I CARCINOGENESIS • prototype is endometrioid carcinoma • Bokhmann 1983 • perimenopausal and early postmenopausal years • type 1 and 2 • associated with oestrogen excess (unopposed oestrogens, PCOS), obesity, hypertension, diabetes • useful as broad concept but much overlap • arises on background of atypical hyperplasia (clinical, morphology) and controversy as to • often low grade and early stage which tumours fall into which group- move • hormone receptor positive away from this broad concept • good prognosis • associated with beta catenin, PTEN, KRAS, PIK3CA , ARID1A mutations and microsatellite instability

15 21/07/2017

ENDOMETRIOID ADENOCARCINOMA Type II Endometrial Cancer

• prototype is uterine serous carcinoma • elderly postmenopausal • not associated with oestrogen excess • no association with • precursor lesion not well defined- serous endometrial intraepithelial carcinoma (serous EIC) • by definition high grade • often advanced stage • hormone receptor negative • poor prognosis • associated with p53, PIK3CA and PPP2R1A mutations (serous)

p53

HIGH GRADE ENDOMETRIAL INTEROBSERVER VARIABILITY CARCINOMAS • grade 3 endometrioid, serous, clear cell, • AJSP 2013;37;874-881- Gilks, Oliva, Soslow carcinosarcoma, undifferentiated, mixed • 56 cases diagnosed as “high grade” endometrial carcinoma • significant interobserver variability in classification • 62% - agreement between all 3 (classification of endometrial carcinomas much more • 36% - major diasgreement (subtype or even whether high problematic than classification of ovarian grade carcinoma was present) carcinomas) • no consistent pattern between observers • major problems- serous v clear cell; serous v grade 3 endometrioid; serous versus undifferentiated • need for molecular tools and development of biomarkers

16 21/07/2017

WHY IS IT SO DIFFICULT TO TYPE TCGA Molecular Classification of ENDOMETRIAL CARCINOMAS? Endometrial Cancers • mixed tumours (possibly more common than in • COPY NUMBER LOW (INTERMEDIATE ovary but probably being over-diagnosed) PROGNOSIS) (39%) • morphologically ambiguous tumours • MICROSATELLITE INSTABILITY HYPERMUTATED • no good marker (WT1 in ovary) (INTERMEDIATE PROGNOSIS) (28%) • COPY NUMBER HIGH (TP53 mutations) (mostly, but not all, serous) (WORST PROGNOSIS) (26%) • POLE ULTRAMUTATED (BEST PROGNOSIS) (7%)

Mutation Spectra Across Endometrial Carcinomas TYPING OF ENDOMETRIAL CANCERS

• TCGA classification of endometrial cancers is of prognostic significance • Tumour histotyping is not (because of poor interobserver reproducibility) • Morphology is not reliable in separating into TCGA categories

Getz et al Nature 2013; 497: 67-73

ProMisE (Proactive Molecular Classsification STATEGIES FOR TYPING (TCGA) Tool for Endometrial Cancers) ‒ Stratified (N=143) endometrial carcinomas into 4 TGCA molecular groups ‒ POLE exonuclease domain mutations • ‒ Mismatch repair protein IHC POLE mutation analysis- identifies POLE ‒ p53 IHC ULTRAMUTATED

• MMR staining (MSH2 and MSH6)- identifies n=143 MICROSATELLITE INSTABILITY HYPERMUTATED “Simplified TCGA • p53 staining- mutation-type staining identifies classifier” MSI (n=41) n=102 COPY NUMBER HIGH • Left with COPY NUMBER LOW n=88 POLE (n=12) • Important for prognosis, management/ targeted therapies CN-H (n=25) CN-L (n=63)

Talhouk A, et al. Br J Cancer. 2015 Jul 14;113(2):299-310

17 21/07/2017

UNDIFFERENTIATED ENDOMETRIAL UK HISTOPATHOLOGY LABORATORIES CARCINOMA

• Need to be developing POLE testing • may be associated with low grade endometrioid • ? Individual laboratories, ? Centralised testing adenocarcinoma (dedifferentiated endometrioid • ? Can we develop surrogate markers of POLE adenocarcinoma or mixed endometrioid and mutation undifferentiated carcinoma)

• similar tumours in ovary

• VERY POOR PROGNOSIS

18 21/07/2017

DEDIFFERENTIATED CARCINOMA IMMUNOHISTOCHEMISTRY

• usually positive for keratins and EMA but staining may be very focal, although strong; EMA staining more consistent than cytokeratins • minor neuroendocrine marker positivity not uncommon (10%) • often abnormalities in MMR protein expression • loss of ER, PR and PAX8 in undifferentiated/ dedifferentiated areas • usually wild-type p53 • loss of E-cadherin in undifferentiated/ dedifferentiated areas

EMA MOLECULAR PATHOLOGY OF UNDIFFERENTIATED/ DEDIFFERENTIATED CARCINOMA • Not studied in TCGA (also clear cell carcinomas and carcinosarcomas) • Probably don’t fit into ProMisE algorithm • Up to 50% (? True (pure) undifferentiated/ dedifferentiated carcinomas) associated with mutations in SWI/ SNF pathway with resultant loss of immunohistochemical staining (SMARCA4, SMARCA2, SMARCB1, ARID) • Also high incidence of MMR abnormalities

SMALL CELL CARCINOMA OF OVARY LOSS OF SMARCA4 (BRG1) OF HYPERCALCAEMIC TYPE- SCCOHT • usually young females (peak in 2nd and 3rd decades) • may occur in older females (large series of 150 cases; AJSP 1994;18;1102-1116; age 9-43; average 24 years) • hypercalcaemia in two-thirds • usually unilateral ovarian tumour • rare familial cases (may be bilateral) • extremely aggressive with poor prognosis

19 21/07/2017

Small Cell Carcinoma of Hypercalcaemic Type- MISNOMER • Small- cells not always small • Carcinoma- no proof that epithelial neoplasm • Hypercalcaemic- serum calcium only elevated in 2/3 cases

RECENT DEVELOPMENTS HISTOGENESIS- SCCOHT • 3 papers in Nature Genetics 2014 • SMARCA4/BRG1 mutated in about 98% of SCCOHT (somatic or • unknown (WHO 2014- categorise among germline mutations) (up to 40-50% have germline) (others miscellaneous ovarian neoplasms) mostly have SMARCB1 mutations) • epithelial, germ cell, sex cord and • SMARCA4/BRG1 (2%) and SMARCB1/INI1 (98%) mutated in virtually all malignant rhabdoid tumours and atypical teratoid/ neuroendocrine origin suggested rhabdoid tumours of CNS • never (until recently) described in association • SMARCA4 and SMARCB1 are core members of SWI/ SNF with another neoplasm complex

20 21/07/2017

IS SCCOHT A FORM OF MALIGNANT RHABDOID TUMOUR? • Clinical similarities (young, aggressive neoplasms) • Morphological similarities (rhabdoid and non- rhabdoid cells) • Molecular similarities (SMARCA4, SMARCB1) • SCCOHT is a form of malignant rhabdoid tumour (could be renamed malignant rhabdoid tumour of the ovary) (but SCCOHT terminology is well established)

SMARCA4 (BRG1) IMMUNOHISTOCHEMISTRY • Loss of nuclear staining in SCCOHT secondary to mutation • Very useful marker of SCCOHT (no useful marker previously) • Retention of nuclear staining in almost all mimics

SMARCA4/ BRG1 DIAGNOSTIC/GENETIC TESTING IMMUNOHISTOCHEMISTRY ALGORITHM FOR SCCOHT • Have the tumour reviewed by a gynaecological pathologist and the diagnosis confirmed (this may necessitate SMARCA4/ BRG1 immunohistochemical staining) • Once the diagnosis is confirmed, counsel the patient for germline SMARCA4 sequencing • In the case of a germline mutation, consider removal of the unaffected ovary in unilateral SCCOHT if only one ovary has been removed • In the case of a germline mutation, test family members for mutation ( ? Role of prophylactic oophorectomy) • In the case of no germline mutation, if appropriate according to age and clinical context, retain unaffected ovary in unilateral SCCOHT

21 21/07/2017

HISTOPATHOLOGY 2017; 70; 1147- CASE ASSOCIATED WITH 1154- PMID 28130795 • 2 cases of SCCOHT in association with • Several similar cases reported in literature • SUGGESTS THAT SCCOHT MAY BE A PRIMITIVE GERM CELL TUMOUR ARISING FROM A TERATOMA

GROWING NUMBER OF TUMOURS The BDIAP Kristin Henry Lecture 2017 WITH SMARCA4 LOSS • SCCOHT • Undifferentiated/ dedifferentiated carcinomas endometrium and ovary • Similar lung, urothelial tract, sinonasal and GI tumours • ? ARE THESE RHABDOID TUMOURS (dyscohesive, sometimes rhabdoid, limited expression of epithelial markers) (? Are dedifferentiated endometrial carcinomas really funny carcinosarcomas)

22