Follicle center lymphoma (FCL) of the lower female genital tract (LFGT): A novel variant of primary cutaneous follicle center lymphoma (PCFCL)

Abstract

Primary cutaneous follicle center lymphoma (PCFCL) has been distinguished from nodal follicular lymphoma (FL) based on genomic and clinical features. The nature of other extranodal FLs is not well defined. We report 15 cases of follicle center lymphoma involving the lower female genital tract (LFGT). Cases were evaluated using an immunohistochemical (IHC) panel for B-cell lymphoma, B-cell clonality, fluorescence in situ hybridization (FISH) for BCL2-gene rearrangement (BCL2-R), and next generation sequencing (NGS). All patients had localized disease with no evidence of bone marrow involvement. Most cases (12/15, 80%) had a follicular pattern, at least focally. Large centrocytes were a prominent feature leading to concern for diffuse large B-cell lymphoma (DLB by referring pathologists. Neoplastic cells were positive for CD20 and BCL-6, while BCL-2 was positive in 2/15 (13%) cases. FISH for BCL2-R was negative in 10/11 (91%) cases. NGS performed in 10 cases revealed TNFRSF14 as the most frequently mutated gene in 6/10 (60%) cases. No case had CREBBP or KMT2D mutations as seen in nodal FL. None of the patients had progressive disease with durable complete remission achieved in 10/12 (83%) cases. The median follow-up period was 7.8 years (range: 0.2-20.5 years) with a 5-year overall survival of 100%. We conclude that FCL of the LFGT is a novel variant of PCFCL. Despite a frequent component of large cells, it is characterized by localized disease and low risk for dissemination. Awareness and recognition are important to distinguish these lesions from aggressive B-cell lymphomas.

Introduction

Follicular lymphoma (FL) is histologically defined as a B-cell neoplasm derived from follicle center cells, usually arranged in a follicular growth pattern, and associated with follicular dendritic cells (FDC). The classical form of nodal FL is a disease of adults, with most patients presenting in advanced stage (Stage III-IVA). It is an indolent and typically incurable disease with multiple relapses. (1) The genetic hallmark of FL is the t(14;18)(q32;q21) translocation. (2) However, secondary alterations are required for FL lymphomagenesis, and mutations in chromatin-modifying genes, including CREBBP, KMT2D, EZH2, and MEF2B, are considered hallmarks of FL. (1, 3, 4)

FL is predominantly nodal, and the involvement of extranodal sites is generally a late event. Some follicle center derived neoplasms are biologically distinct, including primary cutaneous follicle center lymphoma (PCFCL), long recognized by the WHO classification as a distinct entity. (5) It has been distinguished from nodal follicular lymphoma (FL) based on the absence of BCL2 rearrangement (BCL2-R) and the tendency to remain confined to the skin without dissemination. It often contains a predominance of large centrocytes and presents with localized skin lesions on the head or upper trunk. It responds well to local conservative therapy and has an excellent prognosis. (6) By definition, it occurs in the absence of nodal or other extracutaneous involvement at the time of diagnosis. (7) Most cases are negative for BCL-2 protein or show only weak staining; the lymphoid cells rarely show BCL2-R . (5, 8) Recent studies demonstrated that PCFCL differs in its molecular profile from nodal FL in other aspects, besides the absence of BCL2-R. Lesions rarely exhibit mutations in CREBBP and KMT2D, with TNFRSF14 being the most commonly mutated gene.(7, 9) Other primary extranodal sites for FL have been described, including ocular adnexa, salivary glands, and thyroid, but these cases have been not extensively studied for their genomic profile. (10-13)

The most common subtype of lymphoma involving the female genital tract is diffuse large B-cell lymphoma (DLBCL). (14) Most cases occur as secondary spread and are associated with disseminated disease with frequent ovarian involvement. (14-16) Primary lymphomas of the female genital tract are rare and commonly affect the adnexa and cervix, less commonly the uterine body and vagina, and rarely the vulva.(17, 18)

In our consultation practice service, we encountered cases of primary lymphoma involving the cervix and the vagina that were of follicle center derivation, and this led us to search for cases of FL involving the lower female genital tract (uterus, uterine cervix, and vagina) in our archives. The cases identified involved the cervix and vagina, with no cases involving the uterine corpus. We undertook further studies to characterize these lesions in relation to other B-cell lymphomas with particular emphasis on a comparison with nodal and extranodal FL subtypes.

Materials and methods

Selection of Cases

The pathology archives of the National Cancer Institute (NCI), National Institutes of Health, were searched between 01/01/2000 to 05/15/2022 for cases of FL involving the uterus, uterine cervix, and vagina. The slides were reviewed by ESJ and AS. The original submitting pathology reports, molecular and FISH reports, clinical history, treatment history, and clinical outcome/ follow-up were collected from the clinical recordwhenever possible. The study was conducted under an NCI IRB-approved protocol (10CN074).

Histology and Immunohistochemistry

After routine processing, hematoxylin and eosin staining were performed on formalin-fixed, paraffin-embedded (FFPE) tissue sections. Immunostaining on paraffin sections was done for CD20, CD3, Ki67, BCL-6, CD10, CD21, BCL-2, MUM-1, kappa, and lambda (see Table, Supplemental Digital Content 1, which demonstrates antibody specifics). Immunohistochemical stains were performed on a BenchMark ULTRA automated immunostainer (Roche Diagnostics Corporation, Indianapolis, IN) using procedures according to the manufacturer’s instructions. Growth pattern was classified as purely follicular (F), >75% follicular; follicular and diffuse (F & D), 25-75% follicular; focally follicular (FF), <25% follicular; and diffuse (D). The proliferation rate was assessed on Ki67 immunohistochemistry using the percentage of positive cells counted in three high-power fields (400x). The proliferation rate was graded as low, <25%; moderate, 25%-75%; and high, >=75%.

Fluorescence in situ hybridization (FISH)

Fluorescence in situ hybridization (FISH) was performed in 7 cases with available material using BCL2 (Vysis LSI Dual Color BCL2 break apart probe, Abbott Molecular, Inc.) probe and in 8 cases with available material using BCL6 (Vysis LSI Dual Color BCL6 break apart probe, Abbott Molecular, Inc.) probe. The FISH procedure was carried out on FFPE tissue sections, following the manufacturer’s guidelines. After overnight hybridization and subsequent washing, the slides were analyzed using the BioView Duet automated imaging system (BioView, Billerica, MA). At least 100 intact interphase nuclei per case were scored. A cutoff threshold of more than 10% positive cells was used to confirm the presence of BCL2 and BCL6 gene rearrangements. FISH on 4 cases was performed at outside laboratories, and the results were provided by referring physicians.

Fi

Molecular (Clonality and NGS) Studies

DNA was extracted from FFPE sections using the QIAamp FFPE Tissue Kit. Immunoglobulin (IGH and IGK) gene rearrangement studies were performed on 8 cases as previously described. (19) Gene rearrangement studies in 4 cases were performed at outside laboratories, and the results were provided by referring physicians.

In 10 cases, DNA was available for next-generation sequencing (NGS) using the TruSight Oncology 500 panel (see Table, Supplemental digital content 2, which shows genes included in the TSO500 DNA panel) and analyzed using in house genomic data analysis platform called Oncogenomics (NIH, Bethesda, MD). As described previously , a set of parameters was used to filter variants and call them significant.(20) Copy number variation events were reported with detected fold change less than 0.5 for deletion or more than 2.5 for duplication.

Graphics and Data visualization

Several bioinformatics software and packages were used to visualize data and generate graphics. (See word document, Supplemental digital content 3, which shows details of graphics and data visualization)

Overall survival was calculated from diagnosis date until death or last known alive date. Relapse-free survival was calculated from the date of diagnosis until disease progression, relapse, death due to disease (if no earlier disease evaluation was available), or last follow-up evaluation for this disease by the physician. Overall survival and relapse-free survival were estimated using the Kaplan–Meier method. (21)

Results

Clinical Features

The demographic and clinical information are summarized in Table 1. The patients included 15 females with a median age of 41 years (range 31-53). The majority (9/15, 60%) presented with symptomatic mass lesions, with bulky disease (maximum size of lesion > 6 cm) in 5/9 (56%) patients. For the patients presenting with mass lesions (9), the mass involved both cervix and vagina in 3, the vagina in 4, and the cervix in 2 patients. In 8/9 (89%) cases, the mass was limited to the lower female genital tract (cervix and vagina) by clinical examination or imaging with no evidence of involvement of uterine corpus and adnexa. One patient (Case 15) had atypical findings radiologically within the pelvis, but histologically the lesion was confined to the cervix upon supracervical hysterectomy with bilateral salpingo-oophorectomy.

Table 1.

Clinical features of Follicle center lymphoma of the lower female genital tract

Case No.	Age (years)	Site	Clinical presentation	Diagnostic procedure	Submitting diagnosis	Treatment (Rx)	Rx response	Outcome	Follow- up time (years)
1	44	C	menorrhagia	vaginal hysterectomy	DLBCL	Hysterectomy & observation	CR	Alive, NED	0.85
2	43	C	mass lesion	total hysterectomy	Low-grade B cell lymphoma, NOS	NA	NA	NA	NA
3	44	V	mass lesion	vaginal biopsy	FL grade 3 A	RICE x 3 & consolidative RT	CR	Alive, NED	4.2
4	40	C & V	mass lesion, pelvic pain, bleeding	endocervix, endometrium, cervix, vagina biopsy	DLBCL	R-CHOP x 6 & pelvic radiation	CR/RI	Alive, local recurrence by imaging only	4.5
5	36	C	routine screening	endocervix curettage, cervix LEEP	DLBCL	LEEP & R (single agent)	CR	Alive, NED	4.8
6	36	C & V	pelvic pain, abnormal bleeding, dyspareunia	simple hysterectomy	DLBCL	Hysterectomy & observation	CR	Alive, NED	6.9
7	31	C & V	mass lesion, abnormal bleeding	cervix biopsy	FL grade 2	DA-EPOCH-R x 6 cycles	CR	Alive, NED	9.2
8	48	C	gynecological symptoms	hysterectomy	DLBCL	Hysterectomy & observation	CR	Alive, NED	17
9	36	V	mass lesion	vaginal excision biopsy	B cell lymphoma, favor small lymphocytic lymphoma	Brachytherapy	CR	Alive, NED	20.5
10	50	V	mass lesion	partial colpectomy	DLBCL	Observation	CR	Alive, NED	0.22
11	53	V	mass lesion	vaginal biopsy	Sclerosing lymphoproliferative process	NA	NA	NA	NA
12	46	C	mass lesion, abnormal bleeding 2 years post hysterectomy	Trachelectomy	DLBCL	NA	CR	Alive, NED	7.8
13	40	V	routine screening	anterior vaginal wall excision	DLBCL	R-CHOP X 4, RCOP x 1 & involved field RT	CR	Alive, NED	16.5
14	37	C	abnormal bleeding	conization	FL grade 3 with areas of DLBCL	Hysterectomy, CHOP x 5 & R x 18	CR	Alive, NED	13
15	41	C	mass lesion	cervical biopsy and curettings	B cell lymphoma, NOS	Hysterectomy, RCVP, R-CHOP	CR/RI	Alive, NED	9.6

Abbreviations: C, cervix; V, vagina; DLBCL, diffuse large B-cell lymphoma; FL, follicular lymphoma; NOS, not otherwise specified; NA, not available; CR, complete remission; CR/RI, achieved complete remission followed by relapse on imaging; RT, radiotherapy; NED, no evidence of disease; LEEP, loop electrosurgical excision procedure; R, Rituximab; RICE, Rituximab, fractionated Ifosfamide, Carboplatin, Etoposide; R-CHOP, Rituximab, Cyclophosphamide, Doxorubicin hydrochloride (hydroxydaunorubicin), Vincristine sulfate (Oncovin), Prednisone; DA-EPOCH-R, dose-adjusted Etoposide, Prednisone, Vincristine, Cyclophosphamide, Doxorubicin, Rituximab; RCVP, Rituximab, Cyclophosphamide, Vincristine sulfate, Prednisone

4/15 (27%) patients presented with non-specific gynecological symptoms, including abnormal bleeding and pelvic pain with an atypical lymphoid infiltrate identified as an incidental finding in the surgical specimens. 2/15 (13%) patients were asymptomatic and had disease detected at routine gynecological examination (Cases 5, 13).

One patient (Case 3) had a past medical history of large cell lymphoma (of uncertain type) in the vagina diagnosed 19 years ago (pathology report and slides were not available for review), who had achieved complete remission post six cycles of R-CHOP (Rituximab, Cyclophosphamide, Doxorubicin hydrochloride (hydroxydaunorubicin), Vincristine sulfate (Oncovin), and Prednisone). A recurrent vaginal mass lesion led to a biopsy. There was no prior history of lymphoma in any other patients.

All patients had localized disease, negative for marrow involvement (12/12). None of the patients (0/13) had clinical evidence of any other lymphadenopathy or organomegaly. B-symptoms (night sweats) were rarely present, seen only in 1 case (Case 11).

Histological and Immunohistochemical features

All cases were submitted to our laboratory in consultation. The submitting diagnoses were diffuse large B-cell lymphoma (DLBCL), 9/15 (60%); B cell lymphoma, not otherwise specified, 2/15 (13%); FL grade 3, 1/15(7%); FL grade 1-2, 1/15 (7%); sclerosing lymphoproliferative process,1/15 (7%); and B cell lymphoma, favor small lymphocytic lymphoma, 1/15 (7%) (Table 1).

Diagnostic biopsies were submitted from the lower female genital tract, including 7/15 (47%) cervix, 5/15 (33%) vagina, and 3/15 (20%) cervix and vagina. None of the cases with evaluable endometrium (0/7) showed involvement of endometrium or uterine corpus, with involvement being limited to the cervix and vagina.

The cervical and vaginal biopsies showed similar morphologic and immunophenotypic features (Table 2, Figures 1-3). A subepithelial atypical lymphoid infiltrate was present with intact overlying epithelium. None of the cases showed surface erosion or ulceration. A follicular component was seen in 12/15 (80%) cases, at least focally, while 3/15 (20%) cases had a purely diffuse pattern (cases 3, 7, and 15). The infiltrate in 6/15 (40 %) patients exhibited a finely sclerotic background, and 5/15 (33%) biopsies showed perivascular distribution of lymphoid cells. Large centrocytes were a prominent feature of the atypical infiltrate, admixed with other follicle center cells, including centroblasts and small centrocytes in varying proportions. In 8/15 (53%) cases, large centrocytes and centroblasts predominated, while 7/15 (47%) cases contained a predominance of small centrocytes. Centrocytes were the dominant cell type in all cases with a diffuse growth pattern. Mitotic figures were uncommon, seen in only two cases (cases 4, 7).

Table 2:

Pathologic features of Follicle center lymphoma of the lower female genital tract

	Histological Features			Immunohistochemistry, PCR and FISH
Case number	Pattern	Sclerotic background	Predominant cell type	CD20	CD10	BCL-6	BCL-2	CD21	Ki67	IG PCR	BCL2-R	BCL6-R	Copy number changes
1	F	Present	large centrocytes & centroblasts	Pos	Neg	Pos	Neg	Pos Focal	Mod-High	Clonal	Neg	Neg	Neg
2	FF	Present	small centrocytes	Pos	Neg	Pos Focal	Neg	Neg	Low	Clonal	Neg	Neg	Neg
3	D	Absent	small centrocytes	Pos	Neg	Pos	Pos Focal	Neg	ND	Clonal	Neg	Neg	Neg
4	FF	Absent	small centrocytes	Pos	Pos	Pos	Neg	Neg	Mod	Clonal	Neg	Pos	BCL2
5	F	Absent	large centrocytes & centroblasts	Pos	Neg	Pos	Neg	Pos Focal	Mod	Clonal	Neg	Neg	BCL6
6	F & D	Absent	small centrocytes	Pos	Neg	Pos	Neg	Neg	Mod	PC	Neg	Pos	BCL6
7	D	Absent	small centrocytes	Pos	Pos Focal	Pos	Neg	ND	Low-mod	Clonal	Neg	Neg	Trisomy 14
8	F	Absent	large centrocytes & centroblasts	Pos	Pos	Pos	Neg	Pos	Mod - high	IND	Neg	Neg	Neg
9	F	Absent	large centrocytes & centroblasts	Pos	Neg	ND	Neg	ND	ND	Clonal	Neg	Neg	BCL2
10	F & D	Present	large centrocytes & centroblasts	Pos	Neg	Pos	Neg	Pos	Low	PC	Pos	Pos	Neg
11	F & D	Absent	small centrocytes	Pos	Pos	Pos	Neg	ND	ND	INC	ND	Neg	Neg
12	F & D	Present	large centrocytes & centroblasts	Pos	Neg	Pos Focal	Neg	Pos Focal	Low-mod	ND	ND	ND	ND
13	FF	Present	large centrocytes & centroblasts	Pos	Neg	Pos	Neg	ND	ND	ND	ND	ND	ND
14	F & D	Absent	large centrocytes & centroblasts	Pos	ND	Pos	Pos	ND	ND	ND	ND	ND	ND
15	D	Present	small centrocytes	Pos	Neg	Pos	Neg	ND	Low	Clonal	Neg	Neg	Neg

Abbreviations: F, follicular; FF, focally follicular; D, diffuse; F&D, follicular and diffuse; INC, incomplete study (see results); IND, Indeterminate, peak identified only with IgK probe

Figure 1. FCL of the LFGT in a 44-year-old female with menorrhagia for three years underwent hysterectomy (case 1).

(A) Sections of cervix show a submucosal lymphoid infiltrate with ill-defined follicles and intact overlying squamous mucosa. (B) Loosely arranged large lymphoid cells consistent with centrocytes and centroblasts cluster with a perivascular distribution. (C) CD20 is positive in atypical cells and highlights a follicular pattern. (D) CD3 shows T-cells with a perifollicular distribution. (E) BCL-6 is positive in atypical cells, but negative for BCL-2 (F).

Figure 3. FCL of the LFGT in a 50-year-old female who presented with vaginal wall thickening and underwent a partial colpectomy (case 10).

(A) The atypical lymphoid infiltrate has a vaguely nodular pattern in a markedly sclerotic background. (B) Atypical follicular arrangement of lymphoid cells surrounded by sclerosis. (C) CD20 illustrates a follicular and diffuse pattern. (D) CD21 highlights FDC meshworks. (E) Atypical cells show dim staining for BCL-6 and are negative for BCL-2 (F).

The atypical cells were positive for CD20 and BCL-6 in all cases (15/15), while CD10 was positive in 4/14 (29%) cases. Most cases (13/15, 87%) were negative for BCL-2. CD21 was positive in follicular dendritic cells (FDC) in 5/9 (56%) cases. 3 of which had a follicular and diffuse pattern (Case 8, 10, and 12) and 2 cases had a follicular pattern (Case 1 and 5). Most of the cases (8/10, 80%) had a low or moderate proliferation rate and 2 cases (cases 1 and 8) had a moderate to focally high proliferation rate. MUM-1 was performed in 6/15(40%) cases and was negative in all. Kappa and lambda immunostains were performed in 5/15 (33%) cases and were noncontributory, with only interstitial staining and no cytoplasmic positivity.

FISH and Molecular (clonality and NGS) features

FISH for BCL2-R was negative in 10/11 (91%) cases and positive in one case (case 10) (Table 2). FISH for BCL6-R was positive in 3/12 (25%) cases (cases 4,6, and 10). Evidence of copy number increase in non-rearranged BCL2 or BCL6 was seen in 2/11 (18%) and 2/12 (16%)cases respectively.

Immunoglobulin gene rearrangement studies identified a clonal rearrangement pattern in 8/12 (67%) of cases. 2/12 (17%) cases (cases 6 and 10) showed a polyclonal rearrangement pattern but showed pathogenic alterations in EZH2 and TNFRSF14 on NGS. One case(case 11) had an incomplete IG PCR study (only the heavy chain frameworks were performed without IgK tubes) from the submitting lab. When repeat PCR studies were attempted, the quality of the extracted DNA was poor, resulting in no amplification.

NGS (Figures 4, 5) performed in 10 patients revealed TNRFSF14 as the most commonly mutated gene, with mutations in TNFRSF14 found in 6/10 (60%) cases. Other recurrently mutated (>=3 cases) genes included B2M in 4/10 (40%) cases and XIAP in 3/10 (30%) cases. No case had mutations in CREBBP or KMT2D. Two cases showed EZH2 mutations (cases 6, 10), and one case each showed mutations in other chromatin-modifying genes, including HIST1H2BD (case 3), SETD2 (case 7), and TET2 (case 2). One case (case 6) showed copy number variations with duplication of MYC (FC=1.7) and CDK4(FC=1.6) genes.

Figure 4. Mutational landscape of somatic alterations in 10 cases of FCL of the LFGT.

(A) Tumor mutational burden (TMB) showing number of somatic mutations per megabase. (B) Oncoplot: The x-axis denotes each case. The mutated genes are clustered based on the biological pathway on the y-axis. The percentage on the right of oncoplot is the percent of cases in which the gene is mutated. The color coding in the oncoplot is based on the type of alterations, and the samples are annotated with additional properties including immunoglobulin (IG) gene rearrangement by PCR and BCL2 gene rearrangement by FISH. Abbreviations used: Mut: mutation; SNV: single nucleotide variation; UTR5: untranslated region 5; FS: frameshift; Neg: negative; Pos: positive

Figure 5. Lollipop plots of genes recurrently mutated in FCL of the LFGT annotated with oncogenic potential in OncoKb.

The x-axis indicates amino-acid positions. The approximate location of somatic mutations identified in each gene is indicated. The lollipop segment of the mutations identified as oncogenic/predicted oncogenic/likely oncogenic in OncoKb is colored black. The color of the lollipop denotes the type of alteration: green: frameshift substitution, blue: single nucleotide variation, and orange: splicing.

Clinical Course

Follow-up information was available for 13/15 patients for overall survival (OS) and 12/15 patients for relapse-free survival. The median follow-up period (Table 1) was 7.8 years (range: 0.2-20.5 years). The 5-year overall survival was 100% (Figure 6A) and the 5-year relapse-free survival was 76% (Figure 6B).

Figure 6. Survival curves for FCL of the LFGT.

(A) Kaplan Meier curve illustrates overall survival for 13/15 patients. (B). Kaplan Meier curve illustrates relapse free survival for 12/15 patients.

Treatment information was available for 12/15 (80%) cases, and therapeutic approaches varied. 3/12 (25%) patients (Case 1, 6, and 8) had a diagnosis on hysterectomy specimens and had observation and follow-up with no additional therapy. One patient (case 10) was diagnosed on a partial colpectomy specimen and subsequently was followed without further therapy. 4/12 (33%) patients received varied systemic chemo-immunotherapy including RICE (Rituximab, fractionated Ifosfamide, Carboplatin, Etoposide), DA-EPOCH-R (dose-adjusted Etoposide, Prednisone, Vincristine, Cyclophosphamide, Doxorubicin, and Rituximab), RCVP (Rituximab, Cyclophosphamide, Vincristine sulfate, Prednisone), R (single agent) and R-CHOP. 3/12 (25%) patients (cases 3, 4, and 13) received consolidative radiotherapy in addition to chemo-immunotherapy. 1/12 (8%) patient (Case 9) received only brachytherapy.

10/12 (83%) patients achieved durable complete remission following treatment. Two patients had local relapse by imaging; One patient (Case 4) achieved complete remission post six cycles of R-CHOP and pelvic radiation but had a local relapse by imaging (MRI) at the primary site four years after primary diagnosis. Tissue diagnosis was not performed, and the patient was lost to follow-up with no information about further treatment. One patient (case 15) achieved complete remission after multiple forms of systemic therapy and six months later showed local relapse at the primary site by imaging (CT scan). A tissue biopsy was not performed, and the patient was treated with additional systemic chemo-immunotherapy and achieved complete remission. None of the patients were documented to have disease outside of the cervix or vagina.

Discussion

Follicular lymphoma usually presents with widespread nodal disease and only late involvement of extranodal sites beyond the bone marrow. Extranodal FLs are generally distinct from nodal lymphomas clinically and biologically. (12, 13) In this study we report FCL of the LFGT as a distinctive form of B-cell lymphoma presenting in the uterine cervix and vagina. While diffuse areas were frequently present, most cases contained at least focal follicular areas (80%). Large centrocytes and centroblasts were a major component of the infiltrate in 53% of our cases, although small centrocytes were also invariably present. The neoplastic cells exhibited features of follicle center derivation, with all cases positive for CD20 and BCL- 6. CD10 was positive in 29% of the cases, and BCL-2 was positive in only 13%. Most cases (91%) in our study were negative for BCL2-R by FISH. The most common clinical presentation was symptomatic mass lesions in 60% of patients. All cases had localized disease, negative for marrow involvement. Most patients (83%) obtained a durable complete remission following treatment, and none had documented progression of disease outside of the cervix or vagina. However, it should be noted that a number of patients received aggressive combination chemotherapy, which may have impacted their clinical course.

Our findings show that follicle center lymphoma (FCL) of the LFGT is distinct from nodal FL clinically, phenotypically, and at a molecular level. NGS studies identified TNFRSF14 as the most mutated gene in 6/10 (60%) cases, and no case showed CREBBP or KMT2D mutations, which are considered, along with EZH2, the hallmark of nodal FL. (3) Our data provide evidence that FCL of the LFGT is likely a form of PCFCL. Both entities have a prominent component of large centrocytes that may lead to misdiagnosis as DLBCL, which was suspected as the diagnosis in most of our cases. Similar diagnostic pitfalls have been reported in PCFCL. (22-25) Both lesions show diminished CD10 and BCL-2 expression with at least a focally follicular pattern. (9) Like our cases, PCFCL presents with localized disease with absent nodal or other extracutaneous involvement and has an excellent prognosis. (5, 7) In concordance with the molecular findings in our cases, PCFCL has been reported to have significantly fewer mutations in CREBBP and KMT2D, with TNFRSF14 being the most mutated gene. (7, 9)

BCL2-R was absent in the majority (91%) of our cases, similar to the genomic profile of PCFCL. (5, 26) In our study, one case (Case 10) was positive for BCL2-R by FISH. In a recent study by Zhou et al, BCL2-R in PCFCL cases was reported to be predictive of subsequent systemic involvement when combined with mutations in chromatin modifying genes. (7) However, owing to the relatively recent diagnosis of the one positive case, limited time (< 6 months) is available to assess recurrence.

The differential diagnosis of FCL of the LFGT includes lymphoma-like lesions involving the female genital tract reported by Geyer et al., which like our cases, may have a predominance of large lymphoid cells and may have a clonal rearrangement pattern. (27) However, none of our cases had surface involvement or ulceration, which was a prominent feature in the study by Geyer et al. All cases showed a subepithelial band of uninvolved stroma and a deep atypical lymphoid infiltrate with frequent perivascular lymphoid cells and sclerosis. In addition, most of our cases presented with bulky mass lesions in the cervix and vagina, and no case showed involvement of endometrium; in contrast, none of the patients reported by Geyer et al. had a discrete mass, and 25% involved the endometrium.(27)

Other forms of FL that may lack BCL2-R are mainly nodal, including BCL2-R negative, CD23-positive follicle center lymphoma and pediatric type follicular lymphoma (PTFL). (28) BCL2-R negative, CD23-positive follicle center lymphoma typically presents with localized inguinal location, low stage disease, predominant diffuse growth pattern, and is characterized by CD23 expression. (29, 30) Like our cases, BCL2-R negative, CD23-positive follicle center lymphoma has frequent alterations of the TNFRSF14 gene, but in contrast to our data, it is characterized by CREBBP and STAT6 co-mutation, alterations not encountered in our study. (30, 31) PTFL is another entity that, in addition to lack of BCL2-R, is of germinal center origin, has a predominance of large cells, and tendency to remain localized and not recur or progress. (32) PTFL shares with FCL of the LFGT frequent mutations in TNFRSF14 without mutations in BCL2 and epigenetic modifiers, including CREBBP and KMT2D. However, none of our cases contained mutations in the mitogen-activated protein kinase (MAPK) pathway, which has been implicated in a high proportion of cases of PTFL. (33, 34)

Despite the predominance of large cells, most (83%) of the patients in our study achieved durable complete remission following treatment. However, 2/12 (17%) cases experienced local relapse with mass lesions detected at the primary site in post-treatment imaging studies. A tissue diagnosis was not performed in either case. As previously described in mediastinal lymphoma, masses on post-treatment imaging may represent scar tissue, fibrotic change, or necrosis and should not be considered sufficient for further therapeutic intervention in isolation. (35) If relapse is suspected, biopsy for histological confirmation is recommended to guide further management. It is difficult to conclude how these patients should be managed in the future, as 6/15 patients received systemic chemotherapy appropriate for DLBCL. However, 5 patients achieved CR following either local therapy or observation. Further studies are required to determine the optimal management.

We report FCL of the LFGT as a novel form of B-cell lymphoma frequently presenting in the uterine cervix and vagina. We provide evidence for a follicle center derivation, and describe histological, molecular and clinical similarities with PCFCL. Despite a significant component of large cells, it is characterized by localized disease with an apparent low risk for systemic involvement. Recognition of this entity is important to prevent misdiagnosis as DLBCL. Correct diagnosis would probably be challenging in a needle core or punch biopsy. Incisional or excisional biopsy is suggested, and given the characteristic mutational profile, high throughput genomic studies would enhance diagnostic accuracy. In addition, staging to rule out distant sites of disease is further advised.

Supplementary Material

Supplemental Data File 2(.doc, .tif, pdf, etc.)

Supplemental digital content 2. Table which shows genes included in the TSO500 DNA panel. pdf

Supplemental Data File 3 (.doc, .tif, pdf, etc.)

Supplemental digital content 3. Document which shows details of graphics and data visualization. pdf

Supplemental Data File (.doc, .tif, pdf, etc.)

Supplemental Digital Content 1. Table which demonstrates antibody specifics. pdf

Figure 2. FCL of the LFGT in a 36-year-old female with a prominent lymphoid infiltrate on Pap smear underwent endocervical curettage and Loop Electrosurgical Excision Procedure (LEEP) (case 5).

(A) Atypical lymphoid infiltrate surrounds endocervical glands. (B) Large lymphoid cells resemble centroblasts and large centrocytes. (C) CD20 is positive in atypical cells and highlights a focal follicular pattern. (D) CD3 highlights abundant background T lymphocytes. (E) Atypical cells are positive for BCL-6 and negative for BCL-2 (F).

Data access statement:

All genomic data are available on request.