alk-rearranged tumors are highly enriched in the stump ...download.xuebalib.com/3nt4rtvome1y.pdf ·...

ALK-rearranged Tumors are Highly Enriched in the STUMPSubcategory of Uterine Tumors

Kelly A. Devereaux, MD, PhD, Christian A. Kunder, MD, PhD, and Teri A. Longacre, MD

Abstract: Smooth muscle tumor of uncertain malignant potential(STUMP) is a rare diagnosis rendered when there is uncertaintyconcerning the biological potential of a smooth muscle tumor.The initial differential diagnosis is often broad, as tumors inthis subgroup are morphologically heterogenous. Recent datasuggest uterine inflammatory myofibroblastic tumors (IMTs)with anaplastic lymphoma kinase (ALK) rearrangement may bemisclassified as STUMPs, but the extent to which this occurs hasnot been examined. We identified 60 female patients with tumorspreviously diagnosed as STUMP (48 cases) or prospectivelyconsidered for the diagnosis of STUMP (12 cases). Each caseunderwent histologic review, ALK immunohistochemistry (IHC)and confirmatory break-apart fluorescence in situ hybridization(FISH) for ALK if immunoreactive. Six of the 43 (14%) uterineand cervical tumors were ALK IHC positive, whereas tumors atall other sites were ALK IHC negative. Myxoid features, al-though limited in some cases, were present in all 6 ALK IHCpositive tumors, representing 35% (6/17) of tumors displayingmyxoid features at uterine and cervical sites. All ALK im-munoreactive tumors were confirmed to have ALK rearrange-ments by FISH with 1 tumor showing numerous (3 to 8) 3’ ALKsignals, an unusual FISH pattern not previously described inuterine IMTs. Two patients developed recurrent disease and weretreated with ALK-targeted therapy with initial response. Ourdata demonstrate that a significant proportion of uterine andcervical tumors considered to be STUMPs are ALK—positive byIHC and FISH. Future screening of all uterine and cervicalmesenchymal tumors under consideration for the diagnosis ofSTUMP, particularly those with myxoid features, is recom-mended to identify ALK-rearranged IMTs that could potentiallybe treated with targeted therapy using tyrosine kinase inhibitors.

Key Words: smooth muscle tumor of uncertain malignant potential(STUMP), inflammatory myofibroblastic tumor (IMT), myxoidneoplasms, anaplastic lymphoma kinase (ALK)

(Am J Surg Pathol 2018;00:000–000)

U terine smooth muscle tumors of uncertain malignantpotentials (STUMPs) are a rare heterogenous group

of tumors that vary in their histologic appearance,immunohistochemical profile, and clinical outcome. Theterm STUMP was first used by Kempson1 to describeuterine smooth muscle tumors with ambiguous featuresthat precluded an unequivocal classification into a benignor malignant category.2 Diagnostic criteria continue toevolve and can differ based on tumor site and hormonalinfluence; however, the STUMP diagnosis is typically re-served for those uterine or extrauterine female genital tractsmooth muscle tumors in which the type of differentiation(usual, epithelioid, myxoid), manner of necrosis, mitoticindex, degree of atypia, and/or adequacy of sampling ofthe tumor is indeterminate.2–4 Another term that has beenintroduced in the literature for these tumors is atypicalleiomyoma.2 This latter designation has been the source ofconfusion, as this terminology has also been used forleiomyoma with bizarre nuclei, which is a relatively well-defined entity with a low risk of recurrence. Although amajority of patients with uterine STUMPs are cured fol-lowing complete resection, a subset have been found torecur, typically after a prolonged disease-free interval, butoccasionally with a more rapid course.5,6 Predicting whichtumors are at an increased risk for aggressive behaviorcontinues to be a challenge, therefore close long-termfollow-up is recommended for all patients with STUMPsas their biological potential is uncertain.

The challenges associated with predicting clinicalbehavior following the diagnosis of STUMP lie not only inthe reliance on light microscopy to establish whether or notmalignant features are present, but also in the wide variety ofhistologic appearances that broaden the differential diagnosisand cause overlap with several neoplastic processes. Giventhis degree of heterogeneity, it is not surprising that gyneco-logic pathologists differ with respect to their diagnosticthresholds for diagnosing STUMP (or atypical leiomyoma)and the terminology has essentially become a catch-all for aheterogenous collection of female genital tract tumors thatdefy classification. Recognition in recent years that in-flammatory myofibroblastic tumors (IMTs) are more com-mon in the uterus than previously appreciated and are a classof tumor that can easily be confused with a STUMP, weanalyzed 60 tumors that underwent consideration for the di-agnosis of STUMP (48 retrospective and 12 prospective cases)from 2007 to 2017 to determine whether any of these could beclassified as an ALK-rearranged IMT. As both are mesen-chymal tumors with similar histologic and immunohistologic

From the Department of Pathology, Stanford University School ofMedicine, Stanford, CA.

Conflicts of Interest and Source of Funding: The authors have disclosedthat they have no significant relationships with, or financial interestin, any commercial companies pertaining to this article.

Correspondence: Teri A. Longacre, MD, Department of Pathology,L235, Stanford University School of Medicine, 300 Pasteur Drive,Stanford, CA 94305 (e-mail: [email protected]).

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.

ORIGINAL ARTICLE

Am J Surg Pathol � Volume 00, Number 00, ’’ 2018 www.ajsp.com | 1

Copyright r 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

mailto:[email protected]

features, screening for aberrant anaplastic lymphoma kinase(ALK) expression by IHC was performed in each case andconfirmatory break-apart fluorescence in situ hybridization(FISH) was performed if the tumor was ALK immunoreactive.

MATERIALS AND METHODS

CasesFollowing approval by the Institutional Review Board

at Stanford University Hospital in Palo Alto, CA, USA, theStanford pathology database was searched for all STUMPsfrom 2007 to 2017. Forty-eight patients with tumors pre-viously diagnosed as a STUMP that had available formalin-fixed paraffin-embedded (FFPE) tissue were retrospectivelyreviewed. In addition, 12 patients with tumors undergoingfirst time review during this study that were strongly beingconsidered for the diagnosis of STUMP were prospectivelyevaluated. Cases included a total of 60 tumors from femalepatients at pelvic (uterine, n= 41; cervical, n= 2; ovary,n= 1; pelvic, n= 4; vagina, n= 2; peritoneum, n= 1) andnonpelvic (retroperitoneum, n= 3; small bowel, n= 2;abdominal wall, n= 1; vulva, n= 3) sites. Of note, 12 of thecases described above were prospectively reviewed tumorsfrom the uterus (n= 9), cervix (n= 1), pelvis (n= 1), andvagina (n= 1).

IHC StainingIHC studies were performed on 4 μm sections pre-

pared from FFPE tissue using standard autostainingprotocols on the Leica Bond III Instrument (Leica,Buffalo Grove, IL). Specifically, tissue sections weredeparaffinized, rehydrated, and treated with 3% hydrogenperoxide for 15 minutes to quench endogenous peroxidase.Antigen retrieval was performed using bond epitope re-trieval solution (an EDTA-based buffer and surfactant) atpH 9.0 for 30 minutes. The slides were incubated with anALK primary antibody (dilution 1:25, clone 5A4; Abcam,Cambridge, UK), followed by incubation with an anti-mouse secondary antibody conjugated to horseradishperoxidase. Development was performed using a BondPolymer Refine Detection system (Leica) and the 3,3’-diaminobenzidine chromogen. Appropriate positive andnegative controls were included and evaluated with thespecimens tested. Any degree of staining was interpretedas positive. Additional IHC studies were performedon ALK-positive tumors using the following primaryantibodies: desmin (DE-R-11, ready to use; Ventana),h-caldesmon (1:25 dilution; Dako), smooth muscle actin(1A4, 1:200 dilution; Cell Marque), CD10 (56C6, 1:80dilution; Leica), ER (SP1, ready to use; Ventana), and PR(1E2, 1:2 dilution; Ventana).

Fluorescence In Situ HybridizationALK FISH studies were performed on 4 μm sections

prepared from FFPE tissue. Slides were deparaffinized withCitroSolv (Fisher Scientific), digested with CytoZyme Sta-bilized Pepsin (SciGene), pretreated with a sodium thio-cyanate solution at 80°C (VP2000 Pretreatment Solution;Abbott Molecular), refixed in 10% buffered formalin, and

dehydrated in an ethanol series. Dried, dehydrated slideswere denatured with a Vysis HYBrite instrument at 80°Cfor 6 minutes and hybridized for 48 hours at 37°C with theALK break-apart probe (Abbott Molecular). Slides werewashed with posthybridization wash buffer (2× SSC/0.3%NP-40) at 73°C for 2 minutes and counterstained withDAPI. Cells were analyzed using an Olympus BX51 mi-croscope with appropriate fluorescent filters and docu-mented using CytoVision imaging software (LeicaBiosystems). Two hundred interphase nuclei were analyzedfor each specimen. Nuclei demonstrating 5’ (green) and 3’(red) signal separation of > 1 to 2 signal widths, or in-dividual 3’ signal(s) without a colocalized 5’ signal wereconsidered to have a positive ALK rearrangement signalpattern. Tumors with > 8% of cells exhibiting rearrange-ment were considered positive.

RESULTSA total of 60 tumors from female patients were an-

alyzed, including 48 tumors previously diagnosed as aSTUMP with available tissue blocks or unstained slidesand 12 prospectively reviewed tumors in which the diag-nosis of STUMP was strongly being considered (Table 1).Six of the 43 (14%) uterine and cervical tumors were ALKIHC and ALK FISH positive, whereas the tumors at allother sites were ALK IHC negative and, therefore, did notundergo FISH.

The ALK-positive tumors were identified in womenages 23 to 57 years old (mean, 37.5 y) that presented ex-clusively with abnormal uterine bleeding. One of the tumorswas diagnosed within 2 months following pregnancy. Alltumors were localized to the uterine corpus or cervix andranged in size from 3.0 to 8.2 cm (mean, 5.2 cm). The tumorsize could not be determined for one of the cases due tomorcellation at the time of hysterectomy. Three tumors, allin premenopausal women, were small, polypoid masses, andclinically thought to be a polyp or prolapsed submucosalleiomyoma. The fourth tumor in a premenopausal womanwas a more expansive, 8.2 cm tumor thought to be intra-mural uterine fibroid radiographically. The 2 post-menopausal women had significantly enlarged uteri of ∼18and 20 weeks size, respectively, and the masses wereclinically thought to be more diffuse, myometrial-based

TABLE 1. Summary of Anatomic Locations for All TumorsAnalyzedSite Number

Pelvic (n= 51)Uterine* 41Cervix 2Other sites† 8 (1 ovary, 4 pelvis, 2 vagina, 1 peritoneum)

Nonpelvic sites (n= 9)Vulva 3Abdomen 3 (2 small bowel, 1 abdominal wall)Retroperitoneum 3

*Six uterine cases also involved the cervix, ovary, pelvis, abdominal wall, and/or adnexa at diagnosis.

†Two pelvic cases also had associated peritoneal or abdominal wall implants.

Devereaux et al Am J Surg Pathol � Volume 00, Number 00, ’’ 2018

2 | www.ajsp.com Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.


processes with benign entities, such as leiomyomas, in thedifferential. Notably, both of the masses in the post-menopausal women were described as having a “soft” con-sistency intraoperatively or radiographically. Additionalclinical features of the ALK-positive tumors are depicted inTable 2.

Histologically, all ALK-positive tumors were com-prised of spindle cells arranged in a whorled, fascicularpattern. Myxoid stroma was identified in all tumors and theextent was inversely correlated with the tumor cellularity.In general, the uterine masses from the premenopausalwomen (cases 1 and 2) were less cellular and demonstratedmore extensive myxoid stroma (Figs. 1A-D). However,case 3 was unique in that it was the largest in size amongthe premenopausal women and demonstrated intermediatecellularity with less conspicuously myxoid areas as well asmoderate atypia (Figs. 1E-F). The cervical mass from thepremenopausal woman (case 4) showed a heterogenousmorphology with some areas appearing more traditionallymyofibroblastic and myxoid, and other areas displayinghyalinization with cords and nests of entrapped spindlecells (Fig. 2). The uterine masses from the postmenopausalwomen (cases 5 and 6), were more cellular withsignificantly less and focal myxoid stroma (Figs. 3, 4). Asthe majority of the cases were myomectomy orpolypectomy specimens, the borders and presence oftumor infiltration could not be accurately assessed dueto the fragmented nature of the specimens. In the solehysterectomy specimen (case 5), the mass clearlydemonstrated infiltration into the myometrium. Atypiawas scored using the criteria established by Bell et al7 anddetermined to be mild to moderate in the ALK-positivetumors. The mitotic rate in all tumors was low tointermediate, ranging from 0 to 9 mitotic figures per10 high-powered fields. A mixed lymphoplasmacyticinflammatory cell infiltrate was present in all cases. Inaddition, case 1 contained a neutrophilic infiltrate and case6 contained both a neutrophilic as well as an eosinophilicinfiltrate. Fifty percent (3/6) of the tumors displayed adiffuse inflammatory infiltrate, whereas in 50% (3/6) of thetumors the inflammatory infiltrate was more clustered orpatchy. The vasculature within each tumor was notable fora predominance of thin-walled vessels, many of which had anelongated and/or ectatic appearance. The cervical case also

contained occasional thick-walled vessels. Histopathologicfindings are summarized in Table 3.

The six ALK-positive tumors showed conspicuous,aberrant ALK expression by IHC in scattered neoplasticspindle cells. Two ALK staining patterns were observed:diffuse granular cytoplasmic staining (3/6, 50%) and dif-fuse granular cytoplasmic staining with perinuclear ac-centuation (3/6, 50%) (Fig. 5). Results, includingpreviously performed IHC for smooth musclemarkers,hormone receptors, and endometrial stroma markers, aresummarized in Table 4. FISH was performed on the sixALK immunoreactive tumors to confirm ALK generearrangement. Four of the tumors (cases 1, 2, 3 and 5)showed a classic ALK rearrangement break-apart patternby FISH, characterized by separate 3’ and 5’ signals, inaddition to a normal unsplit signal pair. One tumor (case4) demonstrated a variant pattern more commonlyobserved in ALK-rearranged non–small cell lung carci-nomas whereby an isolated 3’ ALK signal is present, inaddition to a normal unsplit signal pair.8 The last case(case 6) displayed an unusual fluorescence signal patternwith numerous (3 to 8) isolated 3’ ALK signals seen, inaddition to a normal unsplit signal pair (Fig. 6). In bothof the latter cases, the isolated 3’ ALK signals wereconsidered positive, assuming the kinase containing 3’ endof the ALK gene underwent a productive translocation andthere was subsequent loss of the 5’ end of the gene and/or achromosomal region.

On the basis of morphologic, IHC, and molecularfindings, 3 of the 33 uterine and cervical tumors previouslydiagnosed as “STUMP” were reclassified as IMTs and 3of the 10 uterine and cervical tumors that underwent ini-tial review during this study and were strongly beingconsidered for the STUMP diagnosis were properly clas-sified as an IMT.

DISCUSSIONUterine STUMPs are a heterogenous subgroup of

tumor with varying appearance and potential for morpho-logic overlap with IMTs. Specifically, both are comprised ofa proliferation of spindle cells arranged in a fascicular ar-chitectural pattern with variably myxoid stroma and canshow similar mitotic indices, tumor cell necrosis, infiltrative

TABLE 2. Clinical Features of Patients With Uterine and Cervical ALK-positive Tumors

Case Age (y) Symptoms SiteSize(cm) Procedure

OriginalDiagnosis

ExtrauterineDisease Recurrence

Follow-upInterval

Status atFollow-up

ALK-inhibitorTherapy

1* 23 AUB Corpus 4.5 Myomectomy IMT† No No 6mo NED No2 26 AUB Corpus 3.0 Polypectomy STUMP No No 2 y NED No3 29 AUB Corpus 8.2 Myomectomy IMT† No No 1mo NED No4 37 AUB Cervix 4.0 Polypectomy IMT† No No 1mo NED No5 53 AUB Corpus NA Hysterectomy STUMP No Yes 4.5 y AWD Yes6 57 AUB Corpus 6.2 Myomectomy STUMP Yes Yes 6.5 y AWD Yes

*Associated with pregnancy.†Prospectively diagnosed during this study.AUB indicates abnormal uterine bleeding; AWD, alive with disease; NA, not available; NED, no evidence of disease.

Am J Surg Pathol � Volume 00, Number 00, ’’ 2018 ALK rearrangements in STUMPs

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved. www.ajsp.com | 3


borders, atypia, and inflammatory infiltrates. In addition,IHC for smooth muscle markers, including caldesmon,desmin, and smooth muscle actin can be positive in bothentities. In contrast to STUMPs however, > 50% of IMTsharbor a defining chromosomal rearrangement involvingthe ALK gene on chromosome 2p23, resulting in aberrantALK expression and activation.2,9,10 In this study, we find a

high proportion (14%) of uterine and cervical tumors eitherpreviously diagnosed as a STUMP or strongly considered tobe a STUMP on prospective review are ALK positive byIHC and FISH and are morphologically compatible withIMTs. Importantly, the 3 cases considered for the STUMPdiagnosis that underwent initial review during the studywere likely correctly diagnosed as an IMT given the recent

FIGURE 1. Cases 1 to 3: Uterine ALK-positive tumors in premenopausal women. The uterine masses in cases 1 and 2 exhibitspindle cells arranged in a loose fascicular architectural pattern with abundant myxoid stroma and prominent thin-walled vessels (A).The cells display mild atypia with low nuclear to cytoplasmic ratios and conspicuous nucleoli (B). Both masses show a diffuselymphoplasmacytic infiltrate (C). Scattered neutrophils are also present in case 1 (D). The uterine mass in case 3 contains more cellularareas with less diffuse and abundant myxoid stroma (E) and the cells are notable for moderate atypia (F).




attention of this entity in the literature and the fact that theyunderwent prospective ALK screening by IHC. Interestingly,none of the 17 STUMPs at other nonuterine and noncervicalsites were ALK-positive by IHC.

While recent studies have raised awareness of leio-myomas and leiomyosarcomas as being potential diag-nostic pitfalls for IMTs in the uterus, this is the first,extensive retrospective review of the STUMP subgroup.Before this study, a single case report described a patientdiagnosed with a uterine STUMP with myxoid featuresthat had recurred and on re-review had been reclassified asan ALK-positive IMT.11 Two additional groups analyzeda total of 17 STUMPs for ALK rearrangements by IHC;however, all were ALK negative and all diagnoses re-mained unchanged. Specifically, Haimes et al12 analyzed asingle uterine STUMP with myxoid features and Pickettet al13 analyzed 16 STUMPs of which the site and pres-ence or degree of myxoid change was not detailed. Thereasons for these discrepant findings are not entirely clear,but may lie in the particular ALK antibody clone used,tumor site or in the individual pathologist’s diagnostic

criteria for the STUMP diagnosis. In our cohort of cases,the ALK-positive tumors identified were all uterine orcervical in origin and all displayed myxoid features. Infact, ∼40% of the uterine and cervical STUMPs in ourstudy showed myxoid features and, of those, 35% wereALK positive. Overall, uterine and cervical STUMPs,particularly those with myxoid features, appear to be themost enriched subgroup of tumors with ALK rearrange-ments to date.

Recent literature has hinted at under-recognition ofuterine IMTs due to their previous misdiagnosis as smoothmuscle tumors. In the published literature, at least 2 of 30(6.7%) myxoid leiomyosarcomas,14 1 of 46 (2.2%)leiomyosarcomas,12,13 at least 4 of 1181 (0.34%)leiomyomas,12,13 and 1 of 18 STUMPs (5.6%)11–13 dem-onstrated ALK positivity by IHC and FISH, promptingrevision of the diagnosis to an IMT on re-review. On thebasis of the high proportion (14%) of ALK-rearrangedIMTs in our uterine and cervical STUMP subgroup, werecommend that all tumors at these sites under consid-eration for a diagnosis of STUMP undergo screening for

FIGURE 2. Case 4: Cervical ALK-positive tumor in a premenopausal woman. The cervical mass is morphologically heterogenousand contains fascicular, myofibroblastic appearing areas (A), more prominent myxoid areas (B) as well as hyalinized areas withcords and nests of entrapped spindle cells (C). The inflammatory infiltrate, comprised of a lymphocytes and plasma cells, is patchyoverall (D).




ALK gene rearrangements. None of the tumors classifiedas a STUMP at nonuterine or noncervical locations werefound to be ALK positive; however, a limitation of thisstudy is the smaller number of cases analyzed at these al-ternative sites. Interestingly, fewer (18%) tumors at non-uterine and noncervical sites displayed myxoid features.

Further investigation of the STUMP subgroup at non-uterine or noncervical sites is required in order to make anybroader screening recommendations.

While this study examined mesenchymal tumorsof the female genital tract for ALK translocations, it iswell established that IMTs in other sites may harbor

FIGURE 3. Case 5: Uterine ALK-positive tumor in a postmenopausal woman with an aggressive clinical course. The uterine mass ismore cellular and comprises of fascicularly arranged spindle cells that infiltrate the myometrium (A) and display a whorled patternin areas (B). The spindle cells show mild cytologic atypia (C). The recurrent tumor is also depicted and is nearly histologicallyidentical to the primary in its architecture, cellularity and minimal myxoid stroma (D to F).




alternative translocations (eg, ROS1, PDGFRβ, RET, andNTRK1/3). Whether non-ALK translocations occur inmesenchymal tumors of the female gynecologic tract is notso well established. Similarly, not all IMTs that are di-agnosed on classic histologic features alone have detect-able translocations. In the course of our study, weencountered 1 such case that was ALK negative by FISHyet we still diagnosed as an IMT, which we did not includein this series. We have also encountered what appeared to

be a uterine leiomyosarcoma with usual (ie, nonmyxoid,nonepitheliod) histologic features that harbored an ALKtranslocation. Although it is reasonable to diagnose amyxoid mesenchymal uterine tumor as IMT in the pres-ence of characteristic histologic features, there are cur-rently no data to support this diagnosis in a uterine tumoron the basis of the presence of an ALK translocationalone. Currently, we do not render a diagnosis of uterineIMT unless the tumor appears to exhibit characteristic

FIGURE 4. Case 6: Uterine ALK-positive tumor in a postmenopausal woman with extrauterine extension and an aggressive clinicalcourse. The uterine mass shows a proliferation of tightly packed spindle cells with only focal areas of myxoid stroma (A). Moderateatypia is present and thin-walled vessels predominate, many of which appeared elongated and dilated (B). Focal areas of coag-ulative or tumor cell necrosis are identified (C). The lymphoplasmacytic inflammatory infiltrate is diffuse and in some areasaccompanied by variably dense eosinophils (D) as well as neutrophils (not depicted).

TABLE 3. Morphologic Features of the ALK-positive TumorsInflammatory Infiltrate

Case Architecture Myxoid Border Necrosis AtypiaMitoses/10

HPFInflammatory

Pattern LYPL NEUT EOSVascularPattern

1 Fascicular Yes NA None 1+ 2 Diffuse + + − TNV2 Fascicular Yes NA Infarct type 1+ 1 Diffuse + − − TNV, ECT3 Fascicular Yes NA None 2+ 3 Patchy + − − TNV, ECT4 Fascicular Yes NA None 1+ 1 Diffuse + − − TNV, TCV, ECT5 Fascicular Yes Infiltrative None 1+ 9 Patchy + − − TNV, ECT6 Fascicular Yes NA Coagulative

type2+ 0-1 Diffuse + + + TNV, ECT

Atypia scored as described by Bell et al7: 3+, severe; 2+, moderate; 1+, mild; 0, none.ECT indicates ectatic appearing vessels; EOS, eosinophils; HPF, high-power field (×40); LYPL, lymphoplasmacytic; NA, not applicable (myomectomy or polypectomy

specimen); NEUT, neutrophils; TCV, thick-walled vessels; TNV, thin-wall vessels.




histologic as well as immunohistologic, and/or molecularfeatures of an IMT. In our practice, we diagnose (rare)tumors of the latter example as leiomyosarcoma on histologicgrounds with a note that the translocation is present, shouldit prove useful for targeted therapy. Further studies withfollow-up are needed to delineate the full parameters forthe diagnosis of uterine IMT.

IHC is a cost effective and efficient way to screentumors indirectly for the ALK gene rearrangement. Whilescreening lymphomas for ALK rearrangements using ALK

IHC is widely practiced, screening nonlymphoma tumors,such as IMTs and non–small cell lung carcinomas, by IHChas been limited by the fact that fusion protein expression isless robust and, therefore, risks a higher false negativeresult.15 Recently, alternative ALK antibodies, such as theD5F3 and 5A4 clones, have proven to be more sensitive atdetecting ALK-fusion proteins, making screening by IHC amore feasible option.16–19 While the ALK gene expression isabsent or minimal in non-neoplastic tissue outside of thecentral nervous system, it is important to note that several

FIGURE 5. ALK IHC patterns. Cases 1 to 6 display 2 ALK-positive staining patterns: diffuse granular cytoplasmic staining withoutperinuclear accentuation (cases 2, 3, and 6) (A to B) and diffuse granular cytoplasmic staining with perinuclear accentuation (cases1, 4, and 5) (C to D).

TABLE 4. IHC and FISH Results for ALK-positive TumorsCase Des H-Cald CD10 ER PR ALK IHC ALK FISH ALK FISH Pattern

1* − − + (patchy) 3+, 70% 3+, 70% Granular cytoplasmic, perinuclear Positive Classic2 + − + (patchy) 2+, 90% 2-3+, 75% Granular cytoplasmic Positive Classic3 + + Weak + (focal) 3+, 70% 3+, 70% Granular cytoplasmic Positive Classic4 − + Weak + (focal) 0, 0% 0, 0% Granular cytoplasmic, perinuclear Positive Isolated 3’ signal5 + + + (patchy) 2-3+, 80% 2-3+, 80% Granular cytoplasmic, perinuclear Positive Classic6 + + Weak + (focal) 2+, 50% 2+, 60% Granular cytoplasmic Positive Multiple isolated 3’ signals

*Smooth muscle actin IHC was performed only on case 1 and is positive.Des indicates desmin; H-Cald, h-caldesmon.




other soft tissue tumors can be positive by ALK IHC. Forexample, immunoreactivity for ALK is observed in epi-thelioid fibrous histiocytomas, which also have ALK generearrangements, as well as other non–ALK rearranged tu-mors, such as myxoid leiomyosarcomas, rhabdomyosarco-mas, malignant peripheral nerve sheath tumors; these lattertumors demonstrate aberrant ALK expression by unknownmechanisms that require further exploration.20–23 In ourexperience, IHC can potentially be used as a screening tool,but the morphologic context and confirmation of ALK re-arrangement by FISH is imperative to establish the diag-nosis.

FISH is considered the gold standard, clinically vali-dated method for detecting ALK rearrangements in tumorsand should still be performed on tumors immunoreactive forALK as well as nonimmunoreactive negative tumors inwhich there is high clinical suspicion for an ALK-rearrangement driven process.15 The majority of IMTs aredriven by gene rearrangements in ALK,24–26 however, aminority of IMTs harbor rearrangements in other genes,such as ROS1, PDGFRβ, RET, and NTRK1/3, of unknownprevalence.24,27–29 Therefore, a negative ALK FISH resultdoes not preclude the diagnosis of IMT. When ALKtranslocations occur, the distal, 3’ end of the gene encodingthe intracytoplasmic kinase domain of ALK, fuses to theproximal 5’ end of another gene, resulting in constitutiveexpression of a chimeric ALK protein. NumerousALK-fusion partners have been identified in IMTs, includ-ing IGFBP5, THBS1, FN1, TIMP3, TPM3, TPM4, EML4,CTCL, RANBP2, SEC31A, DES, and DCTN1.11,12,28,30–34

While ALK FISH is highly sensitive for detecting most re-arrangements, false negatives can occur when there is anintrachromosomal ALK gene inversion. For instance, in areported case of an IGFBP5-ALK fusion, an intra-chromosomal inversion occurred resulting in a small 5’ and3’ split that was indiscernable by traditionally used fluo-rescent microscopy.12 Importantly, this case was positive forALK by IHC, arguing the complimentary contribution ofboth modalities to making the diagnosis of IMT.

The diagnostic caveats of IHC and FISH can beovercome using comprehensive genomic profiling assays,such as RNA sequencing–based fusion methods, to providemore definitive support for the diagnosis of IMT. Genomic-based methods offer an unbiased approach by not onlyscreening for ALK rearrangements, but also simultaneouslyscreening for other gene rearrangements (eg, ROS1,PDGFRβ, RET, NTRK1/3) that alternatively occur inan IMT. In addition, the fusion partner and breakpoint canalso be characterized by this method.15,24 Different fusionpartners likely convey a unique biology, which is evidentin the unique chimeric protein subcellular localization pat-tern detected by ALK IHC. So far, no clear-cut phenotypeor outcomes differences have been established for the dif-ferent ALK fusion partners of breakpoints in IMTs; how-ever, this is an important area for future investigation.35–37

Correctly diagnosing ALK-positive IMTs from theoutset and preventing misclassification is critical, giventheir unique biological potential and the availability ofALK inhibitors. While surgical resection is the standard ofcare and typically curative, 25% recur and <2% metasta-size based on a study of extrapulmonary tumors.10

Defining criteria that prognosticates a more aggressiveclinical course is currently ongoing, but associated featureshave been described for uterine IMTs, including olderpatient age (eg, postmenopausal), larger tumor size, co-agulative necrosis, higher mitotic activity, infiltrativeborders, lymphovascular invasion, and severe atypia.21,30

Different degrees of cellularity and myxoid change havebeen associated with an aggressive clinical course. Forinstance, more aggressive tumors in the Parra-Herranet al21 cohort tended to be predominately myxoid and lesscellular, whereas the more aggressive tumors in theBennett et al30 cohort were more cellularly compact andshowed less myxoid stroma. Lastly, histiocytoid or roundcell morphology has been associated with aggressivevariants at abdominal sites, but has yet to be described inIMTs of the uterus.37 When adjuvant therapy is necessary,IMTs have been shown to respond in trials to tyrosine

FIGURE 6. ALK FISH. A, Four cases (cases 1, 2, 3 and 5) demonstrate a classic ALK rearrangement pattern, characterized by split 3'(red) and 5' (green) probe signals >1 to 2 signal widths, in addition to a single normal, unsplit red-green signal pair (yellow). B,Case 4 displays an isolated 3' signal (red) in the absence of a corresponding 5' (green) signal as well as a normal, unsplit red-greensignal pair (yellow). C, Lastly, case 6 shows numerous (3 to 8) isolated single 3' probe signals (red) in the absence of a corre-sponding 5' signal (green) as well as a single normal, unsplit red-green signal pair (yellow).




kinase inhibitors, such as crizotinib, a small-moleculetargeting ALK, MET, and ROS1 tyrosine kinases.11,38,39

In the reported case of the recurrent “STUMP” that wassubsequently revised to an ALK-positive IMT describedabove, the patient received combination treatment withcrizotinib and a multikinase VEGF inhibitor, pazopanib,which provided ongoing partial clinical and radiographicresponse for over 6 months.11 Importantly, based onhealth factors, age, or the extent of disease, resectionmight not always be possible and the patient may benefitfrom ALK-targeted therapy at initial diagnosis.

Four of the ALK-positive cases in our series (cases 1to 4) were identified in premenopausal young women ofchildbearing age. In 3 of the cases, the tumors were small,and clinically described as being polypoid. Specifically, 1case presented shortly after pregnancy in the setting ofabnormal uterine bleeding and was clinically thought to be aprolapsed, submucosal leiomyoma. Two arose in nullipar-ous women and were clinically thought to be a uterine orcervical polyp. Of note, cervical IMTs are extremely rarewith only 2 other cases reported in the literature.40,41 Thefourth case consisted of a larger, 8.2 cm mass which wasradiographically suggestive of an intramural uterine leio-myoma. Besides the fourth case being larger in size, none ofthe tumors displayed any histologic features known to beassociated with a more aggressive clinical course. If re-currence were to occur, however, ALK-targeted therapycould be a fertility sparing option for these women.

In contrast, one of the IMTs that occurred in apostmenopausal woman (case 5) had a more aggressiveclinical course with recurrence. In addition to her olderage, her primary tumor demonstrated features associatedwith aggressive behavior including a higher mitotic index,increased cellularity with less myxoid stroma and an in-filtrative border.21,30 Following hysterectomy, the pa-tient’s tumor recurred 18 months later as a 14 cm pelvicmass and multiple tumor implants to her abdominal walland small bowel. The patient underwent surgical debulk-ing and at this time the pathology was re-evaluated andALK studies were performed. Given the ALK positivity ofthe tumor by IHC and FISH, her original diagnosis wasrevised to IMT. We had the opportunity to review herrecurrent disease pathology at our institution as well.Importantly, the recurrent tumor closely resembled theprimary tumor histologically and displayed a similar mi-totic index of 9 to 10 mitotic figures per high-powered field(Fig. 3). Twenty-four months after her initial debulkingsurgery, her disease progressed again and she was startedon crizotinib. Although crizotinib initially stabilized herdisease for the first 6 months, she required a seconddebulking procedure while continuing on therapy.

The second clinically aggressive case occurred in a57-year-old woman, who was found to have extrauterineextension and nonresectable pelvic spread only 3 monthsafter her initial diagnostic myomectomy procedure. To date,extrauterine extension has been described in only 7 othercases and portends a worse prognosis.21,30,42,43 In addition toreviewing her initial diagnostic myomectomy specimen, wereviewed her hysterectomy specimen. Besides the patient’s

older age and the larger size of her tumor, histologic featurespredictive of aggressive behavior were also present, includingincreased cellularity as well as focal areas of tumor cell ne-crosis. Genetically, the tumor showed a unique FISH re-arrangement pattern demonstrating several isolated 3’ ALKsignals per cell in the absence of a corresponding 5’ signal.While this pattern has not been reported in the uterus to ourknowledge, it likely reflects greater chromosomal instability,including additional chromosomal losses and gains, com-pared with that of a classic ALK-rearrangement pattern. Inaddition, increased copy number of the 3’ end of the ALKgene may also promote a more aggressive phenotype throughenhanced ALK expression and downstream oncogenic sig-naling. Because of incomplete tumor resection and its moreaggressive phenotype, adjuvant therapy was initiated. Thepatient was briefly treated with chemotherapy, but it wasstopped due to intolerance. She was then started on crizotiniband initially demonstrated partial response and stabilizationof her disease during the first 7 months; however, she pro-gressed thereafter. No additional debulking was performed,but instead she was transitioned to brigatinib, a dual inhibitortargeting ALK and mutated EGFR. On brigatinib, she ex-perienced a dramatic clinical and radiographic response withsustained disease stabilization for 5 years.

Both postmenopausal patients undergoing treatmentwith crizotinib initially showed good response, but thendeveloped disease progression. Interestingly, acquired pointmutations have been described in rearranged ALK genes innon–small cell lung carcinomas and IMTs undergoing cri-zotinib therapy.44 Specifically, a F1174L missense mutationhas been described in an IMT treated with crizotinib.45 It isunknown whether either of the patients’ tumors acquiredany resistance mutations while undergoing treatment withcrizotinib, but it is certainly a possibility.

In summary, we find that there is significant mor-phologic and immunohistologic overlap between STUMPSand IMTs, so much so that the STUMP diagnosis is ahighly enriched, if not the most enriched, subgroup forIMTs in the uterus on retrospective review. Overall, theseresults provide data to support prospective ALK IHCscreening of all uterine and cervical tumors being consideredfor a diagnosis of STUMP in order to more accuratelyidentify patients with ALK-positive tumors that could po-tentially benefit from targeted ALK-inhibitor therapy.

ACKNOWLEDGMENTSThe authors thank Ivy Magonon and Ellen Gomulia for

their contribution to the immunohistochemical studies andCharles D. Bangs and Dr Athena Cherry for their contributionto the FISH analysis and interpretation. In addition, the authorsthank Dr Christina Isacson (CellNetix Pathology, Seattle,WA), Dr Sonya Naryshkin (Mercy Health Laboratory,Janesville, WI), Dr Bijayee Shrestha (El Camino Hospital,Mountain View, CA), and Dr Sharon Wu (El CaminoHospital, Mountain View, CA) for their contribution of cases.

REFERENCES1. Kempson RL. Sarcomas and related neoplasms. In: Norris HJ, Hertig AT,

Abell MR, eds. The Uterus. Baltimore, MD: Williams & Wilkins; 1973.




2. Kurman RJ, Carcangiu ML, Herrington CS, et al. World HealthOrganization Classification of Tumours of Female ReproductiveOrgans. Lyon, France: IARC Press; 2014.

3. Soslow RA, Longacre TA. Uterine Pathology. Cambridge, UK:Cambridge University Press; 2012.

4. Miettinen M, Fetsch JF. Evaluation of biological potential ofsmooth muscle tumours. Histopathology. 2006;48:97–105.

5. Guntupalli SR, Ramirez PT, Anderson ML, et al. Uterine smoothmuscle tumor of uncertain malignant potential: a retrospectiveanalysis. Gynecol Oncol. 2009;113:324–326.

6. Ip PP, Cheung AN, Clement PB. Uterine smooth muscle tumors ofuncertain malignant potential (STUMP): a clinicopathologic analysisof 16 cases. Am J Surg Pathol. 2009;33:992–1005.

7. Bell SW, Kempson RL, Hendrickson MR. Problematic uterinesmooth muscle neoplasms. A clinicopathologic study of 213 cases.Am J Surg Pathol. 1994;18:535–558.

8. Salido M, Pijuan L, Martinez-Aviles L, et al. Increased ALK genecopy number and amplification are frequent in non-small cell lungcancer. J Thorac Oncol. 2011;6:21–27.

9. Chiarle R, Voena C, Ambrogio C, et al. The anaplastic lymphomakinase in the pathogenesis of cancer. Nat Rev Cancer. 2008;8:11–23.

10. Coffin CM, Hornick JL, Fletcher CD. Inflammatory myofibroblas-tic tumor: comparison of clinicopathologic, histologic, and immu-nohistochemical features including ALK expression in atypical andaggressive cases. Am J Surg Pathol. 2007;31:509–520.

11. Subbiah V,McMahon C, Patel S, et al. STUMP un“stumped”: anti-tumorresponse to anaplastic lymphoma kinase (ALK) inhibitor based targetedtherapy in uterine inflammatory myofibroblastic tumor with myxoidfeatures harboring DCTN1-ALK fusion. J Hematol Oncol. 2015;8:66.

12. Haimes JD, Stewart CJR, Kudlow BA, et al. Uterine inflammatorymyofibroblastic tumors frequently harbor ALK fusions withIGFBP5 and THBS1. Am J Surg Pathol. 2017;41:773–780.

13. Pickett JL, Chou A, Andrici JA, et al. Inflammatory myofibroblastictumors of the female genital tract are under-recognized: a lowthreshold for ALK immunohistochemistry is required. Am J SurgPathol. 2017;41:1433–1442.

14. Parra-Herran C, Schoolmeester JK, Yuan L, et al. Myxoid leiomyo-sarcoma of the uterus: a clinicopathologic analysis of 30 cases and reviewof the literature with reappraisal of its distinction from other uterinemyxoid mesenchymal neoplasms. Am J Surg Pathol. 2016;40:285–301.

15. Weickhardt AJ, Aisner DL, Franklin WA, et al. Diagnostic assaysfor identification of anaplastic lymphoma kinase-positive non-smallcell lung cancer. Cancer. 2013;119:1467–1477.

16. Houang M, Toon CW, Clarkson A, et al. Reflex ALK immunohis-tochemistry is feasible and highly specific for ALK gene rearrange-ments in lung cancer. Pathology. 2014;46:383–388.

17. Paik JH, Choe G, Kim H, et al. Screening of anaplastic lymphomakinase rearrangement by immunohistochemistry in non-small celllung cancer: correlation with fluorescence in situ hybridization.J Thorac Oncol. 2011;6:466–472.

18. Conklin CM, Craddock KJ, Have C, et al. Immunohistochemistry isa reliable screening tool for identification of ALK rearrangement innon-small-cell lung carcinoma and is antibody dependent. J ThoracOncol. 2013;8:45–51.

19. Selinger CI, Rogers TM, Russell PA, et al. Testing for ALKrearrangement in lung adenocarcinoma: a multicenter comparison ofimmunohistochemistry and fluorescent in situ hybridization. ModPathol. 2013;26:1545–1553.

20. Li XQ, Hisaoka M, Shi DR, et al. Expression of anaplasticlymphoma kinase in soft tissue tumors: an immunohistochemicaland molecular study of 249 cases. Hum Pathol. 2004;35:711–721.

21. Parra-Herran C, Quick CM, Howitt BE, et al. Inflammatorymyofibroblastic tumor of the uterus: clinical and pathologic reviewof 10 cases including a subset with aggressive clinical course. Am JSurg Pathol. 2015;39:157–168.

22. Cessna MH, Zhou H, Sanger WG, et al. Expression of ALK1 andp80 in inflammatory myofibroblastic tumor and its mesenchymalmimics: a study of 135 cases. Mod Pathol. 2002;15:931–938.

23. Doyle LA, Marino-Enriquez A, Fletcher CD, et al. ALK rearrange-ment and overexpression in epithelioid fibrous histiocytoma. ModPathol. 2015;28:904–912.

24. Lovly CM, Gupta A, Lipson D, et al. Inflammatory myofibroblastictumors harbor multiple potentially actionable kinase fusions. CancerDiscov. 2014;4:889–895.

25. Coffin CM, Patel A, Perkins S, et al. ALK1 and p80 expression andchromosomal rearrangements involving 2p23 in inflammatorymyofibroblastic tumor. Mod Pathol. 2001;14:569–576.

26. Fraggetta F, Doglioni C, Scollo P, et al. Uterine inflammatorymyofibroblastic tumor in a 10-year-old girl presenting as polypoidmass. J Clin Oncol. 2015;33:e7–e10.

27. Alassiri AH, Ali RH, Shen Y, et al. ETV6-NTRK3 is expressed in asubset of ALK-negative inflammatory myofibroblastic tumors. Am JSurg Pathol. 2016;40:1051–1061.

28. Antonescu CR, Suurmeijer AJ, Zhang L, et al. Molecular character-ization of inflammatory myofibroblastic tumors with frequent ALKand ROS1 gene fusions and rare novel RET rearrangement. Am JSurg Pathol. 2015;39:957–967.

29. Yamamoto H, Yoshida A, Taguchi K, et al. ALK, ROS1 andNTRK3 gene rearrangements in inflammatory myofibroblastictumours. Histopathology. 2016;69:72–83.

30. Bennett JA, Nardi V, Rouzbahman M, et al. Inflammatorymyofibroblastic tumor of the uterus: a clinicopathological, immu-nohistochemical, and molecular analysis of 13 cases highlightingtheir broad morphologic spectrum. Mod Pathol. 2017;30:1489–1503.

31. Bridge JA, Kanamori M, Ma Z, et al. Fusion of the ALK gene to theclathrin heavy chain gene, CLTC, in inflammatory myofibroblastictumor. Am J Pathol. 2001;159:411–415.

32. Lawrence B, Perez-Atayde A, Hibbard MK, et al. TPM3-ALK andTPM4-ALK oncogenes in inflammatory myofibroblastic tumors.Am J Pathol. 2000;157:377–384.

33. Ma Z, Hill DA, Collins MH, et al. Fusion of ALK to the Ran-binding protein 2 (RANBP2) gene in inflammatory myofibroblastictumor. Genes Chromosomes Cancer. 2003;37:98–105.

34. Panagopoulos I, Nilsson T, Domanski HA, et al. Fusion of the SEC.31L1 and ALK genes in an inflammatory myofibroblastic tumor. IntJ Cancer. 2006;118:1181–1186.

35. Gleason BC, Hornick JL. Inflammatory myofibroblastic tumours:where are we now? J Clin Pathol. 2008;61:428–437.

36. Cook JR, Dehner LP, Collins MH, et al. Anaplastic lymphoma kinase(ALK) expression in the inflammatory myofibroblastic tumor: acomparative immunohistochemical study. Am J Surg Pathol. 2001;25:1364–1371.

37. Marino-Enriquez A, Wang WL, Roy A, et al. Epithelioidinflammatory myofibroblastic sarcoma: an aggressive intra-abdomi-nal variant of inflammatory myofibroblastic tumor with nuclearmembrane or perinuclear ALK. Am J Surg Pathol. 2011;35:135–144.

38. Butrynski JE, D’Adamo DR, Hornick JL, et al. Crizotinib in ALK-rearranged inflammatory myofibroblastic tumor. N Engl J Med.2010;363:1727–1733.

39. Mosse YP, Lim MS, Voss SD, et al. Safety and activity of crizotinibfor paediatric patients with refractory solid tumours or anaplasticlarge-cell lymphoma: a Children’s Oncology Group phase 1consortium study. Lancet Oncol. 2013;14:472–480.

40. Gucer F, Altaner S, Mulayim N, et al. Invasive inflammatorypseudotumor of uterine cervix: a case report. Gynecol Oncol. 2005;98:325–328.

41. Abenoza P, Shek YH, Perrone T. Inflammatory pseudotumor of thecervix. Int J Gynecol Pathol. 1994;13:80–86.

42. Rabban JT, Zaloudek CJ, Shekitka KM, et al. Inflammatorymyofibroblastic tumor of the uterus: a clinicopathologic study of 6cases emphasizing distinction from aggressive mesenchymal tumors.Am J Surg Pathol. 2005;29:1348–1355.

43. Kushnir CL, Gerardi M, Banet N, et al. Extrauterine inflammatorymyofibroblastic tumor: a case report. Gynecol Oncol Case Rep.2013;6:39–41.

44. Ardini E, Galvani A. ALK inhibitors, a pharmaceutical perspective.Front Oncol. 2012;2:17.

45. Sasaki T, Okuda K, Zheng W, et al. The neuroblastoma-associatedF1174L ALK mutation causes resistance to an ALK kinase inhibitorin ALK-translocated cancers. Cancer Res. 2010;70:10038–10043.




本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/

http://www.xuebalib.com/

http://www.xuebalib.com/cloud/


http://www.xuebalib.com/vip.html

http://www.xuebalib.com/db.php

http://www.xuebalib.com/zixun/2014-08-15/44.html


alk-rearranged tumors are highly enriched in the stump ...download.xuebalib.com/3nt4rtvome1y.pdf ·...

Documents