Characterizing Paratesticular Neoplasms in Proteus Syndrome

Abstract

Proteus syndrome is a rare mosaic overgrowth disorder caused by somatic activating variants in AKT1, most commonly the c.49G>A p.(Glu17Lys) variant. It predisposes individuals to asymmetric tissue proliferation and an elevated risk for both benign and malignant neoplasms. Among 64 males with genetically confirmed Proteus syndrome enrolled in a longitudinal natural history study, 12 (19%) underwent surgery for paratesticular masses. The average age at surgery was 9 yrs, most tumors were unilateral, small (median 1.6 cm), and slow growing, but 50% showed recurrence or metachronous tumor development, occasionally with progression to more atypical histology. Histologically, these tumors demonstrated a broad spectrum of differentiation. Eight reviewed cases included Müllerian-type papillary cystadenomas and low-grade papillary adenocarcinomas, a Brenner tumor, and one case of a papillary adenocarcinoma with spindle cell transformation. The epithelial components were typically arranged in papillary and glandular architectures, with variable degrees of cytologic atypia. Psammomatous calcifications were common. Immunohistochemistry showed consistent expression of PAX8(7/7), WT1(7/7), estrogen receptor (ER)(7/7), and progesterone receptor (PR)(6/7), supporting Müllerian lineage, while negative staining for germ cell, and mesothelial markers excluded common paratesticular differential diagnoses. Four of 7 tumors were positive for SF-1. All 7/7 sequenced tumors harbored the AKT1 c.49G>A variant with no additional oncogenic alterations identified by exome sequencing. This series is the largest series to date documenting the clinicopathologic features of paratesticular tumors, a poorly understood component of the Proteus syndrome phenotype.

Introduction

Proteus syndrome is a rare, sporadic, mosaic overgrowth disorder characterized by asymmetric and disproportionate tissue proliferation involving skin, bone, adipose tissue, and various internal organs.¹ The condition is caused by a somatic activating variant in the AKT1 gene (c.49G>A, p.(Glu17Lys)), leading to dysregulated cell growth and differentiation.² While the clinical presentation is highly variable, individuals are at increased risk for developing a wide range of benign and malignant tumors, bony, adipose, and connective tissue overgrowth, and vascular malformations.^{1, 3, 4}

Urologic manifestations of Proteus syndrome are infrequently reported in the literature and are poorly characterized. Paratesticular masses represent a frequent, but less well understood, feature of the disease, with limited data describing their histologic spectrum, clinical behavior, or optimal management. To date, there are only sporadic case reports of paratesticular tumors in Proteus syndrome, and no comprehensive series has been published. As a result, the incidence, pathologic diversity, and potential implications for surveillance and treatment remain largely unknown. Here, we present a series of individuals with Proteus syndrome who developed paratesticular tumors, detailing the clinical presentation, radiologic and pathologic findings, and outcomes. Our aim is to define the spectrum of these tumors, assess their clinical significance, and provide recommendations for their evaluation and management.

Materials and Methods

Participants

This study was conducted as part of an ongoing natural history study of Proteus syndrome (NCT00001403). Inclusion criteria for the current investigation included a clinical or clinical-molecular diagnosis of Proteus syndrome, updated and refined by Sapp et al.^{5, 6} Those with a pathologically confirmed paratesticular tumor underwent chart review. When a tumor sample was available central pathology review was conducted. Individuals with only ultrasonographic evidence of cyst or lesion were not included.

A chart review was conducted to identify the clinical presentation, histologic diagnosis, age at tumor surgery, laterality of tumor, widest gross diameter of tumor as reported by pathology, treatment, recurrence of tumor, occurrence of any other tumors, and mortality status. All living participants were recontacted to inquire about any recurrence of tumor.

Molecular Analysis

Cores of formalin-fixed paraffin-embedded tissue were obtained to increase tumor fraction and improve detection of tumor specific genetic alterations as described.⁷ Tumor and non-tumor (blood, saliva, or adjacent normal) tissue underwent low depth genome sequencing and exome sequencing followed by chromosomal structure, copy number variant, and single nucleotide variant analyses. Genomic DNA was extracted from 2mm needle core punches from FFPE blocks as selected by investigator JKK. The cores were deparaffinized in molecular biology grade mineral oil with a cocktail of ATL buffer and Protinase K (Qiagen) at 7:1 in the Lysis buffer. They were incubated overnight at 65°C and the DNA was purified from the lysate by binding and eluting on a Qiagen Qiaquick spin column. The DNA purity was characterized spectrophotometrically on a Nanodrop, and was quantitated fluorometrically on a Qubit. DNA (200ng) was sheared on a Covaris E220 to ~175bp and the NGS libraries generated with the KAPA Hyperprep. The libraries were pooled and hybridized to the IDT xGen^tm Exome Hyb Panel before they were then run on the Illumina NextSeq 500. Chromosomal structure was analyzed using Nexus Copy Number Software. Somatic variants were identified using Mutect caller.

Tissue samples and immunohistochemistry

Tissue samples from eight tumors from seven individuals were available for histopathology review and immunohistochemistry. Immunohistochemical stains were performed at Johns Hopkins Hospital following standard protocols. The antibodies used, vendors, pretreatments, and dilutions are presented in Table 1. When the proportion of positive tumor cells was <10 % with any intensity, the result is reported “+, focal”, when the proportion of positive tumor cells with any intensity was 10–50%, the result is reported as “+, patchy” and >50% positive tumor cells with any intensity, the result is reported as “+”.

Table 1.

Summary of antibodies and staining methods used for immunohistochemistry

Antibody	Platform	Source	Host	Clone	Dilution	Detection Method
CAIX	Leica Biosystems	Leica Biosystems, (Richmond, IL)	Mouse monoclonal	TH22	1:100	Bond refine DAB
Calretinin	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN)	Rabbit polyclonal	SP65	predilute	Optiview DAB
Chromogranin	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN)	Mouse monoclonal	LK2H10	predilute	Ultraview DAB
Cytokeratin 7	Benchmark Ultra Roche	Dako (Carpinteria, CA)	Mouse monoclonal	(M7018) OV-TL 12/30	1:500	Ultraview DAB
Cytokeratin 20	Benchmark Ultra Roche	Cell Marque (Rocklin, CA)	Mouse monoclonal	Ks20.8	predilute	Ultraview DAB
ER	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN)	Rabbit monoclonal	SP-1	predilute	Ultraview DAB
GATA3	Benchmark Ultra Roche	Biocare (Austin, TX)	Mouse monoclonal	L50–823	1:100	Ultraview DAB
INSM1	Benchmark Ultra Roche	Santa Cruz Biotechnology (Dallas, TX)	Mouse monoclonal	A8	1:200	Ultraview DAB
Ki-67	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN)	Rabbit monoclonal	30–9	predilute	Ultraview DAB
Napsin-A	Leica Biosystems	Leica Biosystems, (Richmond, IL)	Mouse monoclonal	IP64	1:800	Bond refine DAB
PAX8	Benchmark Ultra Roche	Cell Marque (Rocklin, CA)	Rabbit polyclonal	EP331	Predilute	Ultraview DAB
PR	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN))	Rabbit monoclonal	1E2	predilute	Ultraview DAB
P40	Benchmark Ultra Roche	Biocare (Austin, TX)	Mouse monoclonal	BC28	1:100	Ultraview DAB
P53	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN)	Mouse monoclonal	BP53–11	Predilute	Ultraview DAB
P63	Benchmark Ultra Roche	Biocare (Austin, TX)	Mouse monoclonal	4a4	predilute	Ultraview DAB
RB	Benchmark Ultra Roche	Cell Signaling Technologies (Danvers, MA)	Mouse monoclonal	4H1	1:1500	Optiview DAB
SALL4	Leica Biosystems	Millipore Sigma (Burlington, MA)	Mouse monoclonal	6e3	1:400	Bond refine DAB
SF-1	Leica Biosystems	Thermo Fisher Scientific (Waltham, MA)	Mouse monoclonal	N1665	1:50	Bond Refine DAB
Synaptophysin	Leica Biosystems	Novocastra- Leica Biosystems, (Richmond, IL)	Mouse monoclonal	27G12	1:400	Bond refine DAB
TTF1	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN)	Mouse monoclonal	8G7G3/1	predilute	Ultraview DAB
Uroplakin-2	Benchmark Ultra Roche	Biocare (Austin, TX)	Mouse monoclonal	BC21	1:100	Ultraview DAB
WT1	Benchmark Ultra Roche	Roche Diagnostics (Indianapolis, IN)	Mouse monoclonal	6F-H2	predilute	Ultraview DAB

Results

At the time of review, there were 64 males enrolled in the natural history study of Proteus syndrome. Twelve individuals (19%) had a history of surgery for a paratesticular tumor, and their clinical characteristics are summarized in Table 2.

Table 2.

Clinical characteristics of participants

Participant/Tumor # in this study	Previously published	Original diagnosis	Age at surgery (Years)	Laterality	Tumor size (cm)	Orchiectomy	Follow-up from initial surgery (years)	Second tumor	Interval to 2nd tumor (years)	Second diagnosis	Laterality 2nd tumor	Size of 2nd tumor (cm)	Orchiectomy	Reviewed diagnosis
1/1	No	Carcinoma*†	13	Bilateral	4.5	Yes	2‡	No	NA	NA	NA	NA	NA	Müllerian -type Papillary cystadenoma
2/2&6	Biesecker, 20011	Cystadenoma*	5	Unknown	1.5	No	25	Yes	10	Carcinoma*†	Bilateral	4	Yes	Müllerian -type Papillary cystadenoma / Papillary adenocarcinoma
3/3	Biesecker et al., 202010	Cystadenoma*†	17	Right	4.8	No	9	No	NA	NA	NA	NA	NA	Müllerian -type Papillary cystadenoma
4/4	No	Carcinoma*†	18	Left	1.6	Yes	1‡	No	NA	NA	NA	NA	NA	Low-grade papillary adenocarcinoma
5/5	No	Collagenoma	15	Left	4.5	No	14	Yes	10	Carcinoma*†	Left	9.5	No	Low grade papillary adenocarcinoma
6/7	Sadaf, et al., 20188	Borderline tumor*†	4	Left	0.8	Yes	8	Yes	8	Unknown	Right	NA	NA	Papillary Adenocarcinoma with spindle cell transformation
7/8	No	Cystadenoma	26	Right	Unknown	No	15	Yes	8	Brenner tumor*†	Right	3	Yes	Brenner-like tumor
8/NA	No	Cystadenoma	19	Left	1.1	No	36	No	NA	NA	NA	NA	NA	NA
9/NA	Hornstein et al., 1987 (patient 2)11	Carcinoma	4	Left	2.5	Yes	44	No	NA	NA	NA	NA	NA	NA
10/NA	Buser et al., 20209	Borderline tumor	2	Left	1.2	Yes	6‡	No	NA	NA	NA	NA	NA	NA
11/NA	No	Cystadenoma	2	Left	Unknown	No	25	Yes	15	Unknown	Right	Unknown	No	NA
12/NA	No	Carcinoma	14	Right	2	Yes	17	Yes	3	Borderline tumor	Left	1	Partial	NA

^*sample with histologic review

^†sample with molecular analysis

^‡Deceased of nontumor cause

NA: Not applicable

Nine (75%) were alive and three (25%) were deceased. No deaths were related to a paratesticular tumor. Eleven (92%) had a clinical-molecular diagnosis of AKT1-related Proteus syndrome and one had a clinical diagnosis of Proteus syndrome, without molecular confirmation. Tissue was not available from this individual to confirm an activating mosaic AKT1 variant.

The median age of initial surgery was 13.8 (range 1.8–26.1) years. Four individuals were followed with serial imaging before surgery for a median of 5.0 (range 2.2–10.1) years. Initial tumors were unilateral in 11 individuals occurring on the left in seven individuals (58%), on the right in three individuals (25%), bilateral in one individual (8%), and in one individual (8%) laterality was unknown. The median widest diameter of tumor as measured by gross pathology was 1.6 (range 0.8–4.8) cm. The anatomic locations of the tumors were surgically described as being near the upper pole of the testis including paratestis, epididymis, rete testis, and tunica vaginalis. None of the tumors were intratesticular. Eleven individuals received surgical treatment only and one individual received post-surgical adjuvant chemotherapy. Recurrence or second metachronous paratesticular tumor occurred in six individuals (50%) at a median time to recurrence of 7.1 (range 2.2–14.0) years. Diagnoses at recurrence or second tumor were carcinoma in two, borderline tumor in one, Brenner tumor in one, and unknown in one. One individual, who did not have a recurrent or second paratesticular tumor, developed extragonadal neoplasms including a renal cell carcinoma and meningioma. Five (42%) of the individuals included have previously been reported as case reports or in case series.^{1, 8–11}

Of the 12 individuals with surgery for paratesticular tumor, the initial tumor diagnoses were cystadenoma in five (42%), borderline tumor in two (17%), carcinoma in four (33%), and collagenoma in one (8%). Tumor specimens and genetic sequencing were available for seven tumors from seven individuals. The activating AKT1 c.49G>A p.(Glu17Lys) variant was identified in all seven samples available for assay. Variant allele fractions ranged from 35%−78%. No secondary driver variants were identified. Aneuploidy was identified in one tumor, a Brenner tumor with a chromosome gain of 7q (Supplemental Table).

Histopathology and Immunohistochemistry

Material from eight tumors was available to review and to perform immunohistochemical stains. Each tumor was from a different individual except for one individual who presented with a recurrence ten years following the initial resection. The diagnoses presented here are based on the authors’ review of the material, the original diagnoses rendered at the time of the excision are listed in Table 2.

Tumors 1, 2, and 3 were alike (Figure 1), and consisted of an exophytic papillary tumor arising from the tunica albuginea. The papillary structures were lined by cuboidal epithelium with areas showing a ciliated apical border. There were rare psammomatous calcifications. There was no necrosis, stromal or vascular invasion. Overall, the morphology of these tumors corresponded to a Müllerian type- papillary cystadenoma.

Figure 1. Tumor 1, Müllerian type papillary cystadenoma.

A. Low power view of the base of the stalk of a papillary tumor with exophytic growth originating in the tunica albuginea. A small fragment of testicular parenchyma is attached (*). B, C, D. Low power view of the same tumor as in “A” showing the arborization of the papillary structures. The stroma of some papillary cores, particularly towards the distal tip, show edematous dilatation (C). E&F, Higher magnification shows the epithelial lining of the papillary structures is composed of columnar cells with areas of ciliated apical border (F). A-F: Hematoxylin and eosin stained (H&E). A-C: 20x magnification. D: 40x magnification. E: 100x magnification. F: 200x magnification.

Tumor 4 originated adjacent to the epididymis and consisted of a tumor with prominent compacted papillary structures, with areas of solid growth with nest formation and growing within a cystic structure (Figure 2A). The epithelial lining was heterogenous within the tumor. In areas where the papillae were more open and separated from each other, the lining was cuboidal cells with scant eosinophilic cytoplasm and stratification with distal detached cellular tufts, with some resemblance to serous borderline tumors (Figure 2B). The more compacted papillary areas had an epithelium with columnar cells with elongated nuclei with significant nuclear overlap, and subnuclear clear cytoplasm (Figure 2C). The solid nests within the stroma were composed of cells with atypical nuclei with open chromatin and clear cytoplasm (Figure 2D). There was one larger cyst lined with cuboidal epithelium with small papillary tufting and surrounding cellular, ovarian-like stroma (Figure 2E). Psammomatous calcifications were abundant. Mitoses were inconspicuous and there was no vascular or perineural invasion seen. Taken together, the morphology was of a low-grade papillary adenocarcinoma.

Figure 2. Tumors 4, 5 and 6. Low-grade papillary adenocarcinoma.

A. Low power view of tumor 4 located adjacent to the epididymis (*) and with a predominantly papillary architecture (H&E, 20x magnification). B. Same tumor as in “A” with higher magnification in an area of the tumor with open papillary structures lined by cuboidal epithelium with detached small tufts, similar to a borderline serous tumor (H&E, 100x magnification). C. Same tumor as in “A” with higher magnification in area of low-grade adenocarcinoma with papillary structures lined by cuboidal epithelium with nuclei with open chromatin and conspicuous small nucleoli (inset) (H&E, 100x magnification, inset 400x magnification). D. Same tumor as C with area of tumor nests within stroma (H&E, 100x magnification). E. Tumor 5 with large cyst lined by cuboidal cells and with areas of subepithelial ovarian-like stroma (H&E, 40x magnification). F. Low power view of tumor 6 showing a predominantly solid architecture with areas of cholesterol clefts (H&E, 20x magnification). G. Same tumor as in “F” showing an infiltrative tumor composed of poorly formed angulated glandular structures in desmoplastic stroma (H&E, 400x magnification). H. High power view of tumor 6 showing an area of solid growth with tumor cells with atypical nuclei, nuclear pleomorphism and rare mitotic figures. (H&E, 400x magnification).

Tumor 5 and 6 (recurrence in individual with tumor 5) were morphologically similar (Figure 2F-H). Like tumors 1 and 2, these had areas of papillary architecture but were predominantly solid with irregular infiltration of epithelial cells arranged in nests and cords and forming elongated glandular structures. There was marked nuclear atypia with large pleomorphic nuclei with prominent nucleoli. Rare mitoses were seen. The stroma was desmoplastic with cholesterol clefts and foamy histiocytes. The predominantly solid architecture with desmoplastic stroma and irregular infiltration of the stroma by highly atypical cells were characteristic of an invasive adenocarcinoma.

Tumor 7 also grew within a cyst with a fibrous capsule and was surrounded by soft tissue elements including adipose tissue, thick-walled vessels and small portion of epididymis. The tumor was composed of papillary structures lined by a single layer of cuboidal cells. There were solid areas that turned into packed spindle cell areas with elongated overlapping nuclei with open chromatin and prominent nucleoli (Figure 3). Rare mitoses were seen in the spindled areas. The stroma was hyalinized. There was no vascular invasion. Overall, the findings were consistent with a papillary adenocarcinoma with spindle cell transformation.

Figure 3. Tumor 7. Papillary adenocarcinoma with spindle cell transformation.

A. Low power view shows the interface between the papillary component and the solid growth with spindle cells (H&E, 40x magnification). B. Immunohistochemistry for SF-1 shows diffuse staining in the nuclei of tumor cells, both in the papillary and solid/spindle cell components (100x magnification). C. Immunohistochemistry for PAX8 showing diffuse nuclear staining in papillary and solid/spindle cell areas (40x magnification). D. Immunohistochemistry for WT1 showing diffuse nuclear staining in papillary and solid/spindle cell areas (100x magnification). E. Immunostaining for p53 shows weak to moderate scattered positive tumor cell nuclei in the papillary component (wild-type pattern) and strong diffuse staining in the spindle cell component, likely a mutant pattern (100x magnification). F. Immunostaining for Ki67 show frequent positive nuclei both in the papillary and in the solid components (100x magnification). G. Immunohistochemistry for CK7 shows patchy staining in the papillary component, while the solid spindle cell component is negative (40x magnification). H. Immunostaining for ER show diffuse strong positive nuclei in the papillary component, while diffuse moderate staining in the solid spindle cell component (40x magnification).

Finally, tumor 8 (Figure 4), consisted of a multiloculated cyst with bland epithelial lining resembling transitional epithelium of a Brenner tumor. The stroma was hyalinized. There was no stromal invasion, mitoses, necrosis or lymphovascular (LVI) invasion; although we were not able to perform immunostains for vascular markers to detect foci of LVI potentially missed on H&E-stained slide. A summary of the morphologic features of all the tumors is presented in Table 3.

Figure 4. Tumor 8. Brenner-like tumor.

A. Low power view of a multiloculated tumor with cystic spaces lined by bland transitional epithelium (H&E). B. Immunohistochemistry stain for p63 is diffusely positive in the basal to mid-levels of the epithelium while negative or weak towards the surface. C. Immunohistochemistry for GATA3 shows diffuse nuclear staining or epithelial cells from base to apical layers. D. Immunostaining for Ki67 shows positive nuclei limited to the basal cell layers.

Table 3.

Summary of histopathologic features

Tumor	Diagnosis	Cystic spaces	Architecture	Epithelial cells	Stroma	Psammommatous calcifications	Necrosis	Lymphovascular invasion	Mitoses
1	Müllerian type- papillary cystadenoma.	No	Papillary, growing off the tunica albuginea, edema	Columnar, focally ciliated	Hyalinized, sclerotic, focally edematous	Yes	No	No	No
2 *	Müllerian type- papillary cystadenoma.	Yes	Papillary	Columnar, focally ciliated	Hyalinized	Yes	No	No	No
3	Müllerian type- papillary cystadenoma.	No	Papillary	Columnar epithelium, ciliated	Hyalinized	Yes	No	No	No
4	Low grade papillary adenocarcinoma	Yes	Papillary and solid	Cuboidal in areas of open papillae, columnar in more compacted areas	Hyalinized, desmoplastic, Abundant foamy histiocytes, cholesterol clefts	Yes	Yes, focal	No	Yes
5	Low grade papillary adenocarcinoma	No	Papillary and confluent nests. Invasive solid and poorly formed glandular structures	Cuboidal, clear cytoplasm, marked nuclear atypia with prominent nucleoli.	Fibroelastotic, hyalinized/fibrotic, ovarian-like in large cyst	Yes	Yes	No	Yes, rare
6 *	Low grade papillary adenocarcinoma	Yes	Papillary and confluent nests. Half of the tumor is solid.	Cuboidal, clear cytoplasm, marked nuclear atypia with prominent nucleoli	Desmoplastic, and sclerosed	Yes	No	No	No
7	Papillary carcinoma with spindle cell transformation	Yes	Papillary and solid with spindled areas	Cuboidal in papillary areas, spindled	Hyalinized	Yes	No	No	Yes, in spindle areas
8	Brenner-like tumor	Yes	Cysts and nests with bland transitional epithelium	Stratified transition	Hyalinized	No	No	No	No

^*Recurrence in same individual.

Immunohistochemistry showed that the tumors stained similarly and were characterized by diffuse positive staining for PAX8 and WT1, focal and patchy staining for CK7, frequent staining for ER and PR. p53 showed a wild-type pattern in all cases 1–7, except for the spindle cell areas of tumor 7 where it was overexpressed, suggestive of a mutant pattern (no molecular testing was available for confirmation). SF-1 was positive in epithelial cells of four tumors, two diffuse, and two focal. Neuroendocrine markers, TTF1, p40, Napsin-A, were consistently negative in all tumors. See table 4 for summary of immunohistochemistry results.

Table 4.

Summary immunohistochemistry results

Marker/tumor	1	2	3	4	5	6	7	8
CAIX	+ focal	+ patchy	+ patchy	+ focal	+ patchy	+ patchy	+ patchy	NA
Calretinin	-	+ focal	+ focal	-	+ focal	-	-	NA
Chromogranin	-	-	-	-	-	-	-	NA
Cytokeratin 7	+ focal	+ patchy	+, patchy	+ patchy	+, patchy	+	+ patchy	NA
Cytokeratin 20	NA	NA	NA	NA	NA	NA	NA	-
ER (%, intensity)	>95, strong	10, weak	5, mod	90, strong	5, mod	90, mod	90, strong	NA
GATA3	NA	NA	NA	NA	NA	NA	NA	+
INSM1	-	-	-	-	-	-	-	NA
Ki-67 (%)	10–15	2–5	2–10	2–10	2–10	5	10–20	5–10, basal
Napsin-A	-	-	-	-	-	-	-	NA
PAX8	+	+	+	+	+	+	+	NA
PR (%, intensity)	80, mod	0	5, weak	90, strong	5, weak	50, mod	80, mod	NA
P40	-	-	-	-	-	-	-	NA
P53	WT	WT	WT	WT	WT	WT	WT/Mut in spindled areas	NA
P63	NA	NA	NA	NA	NA	NA	NA	+, basal-mid
RB	Retained	Retained	Retained	Retained	Retained	Retained	Retained	NA
SF1	+ focal	-	-	+ focal	+	-	+	NA
SALL4	-	-	-	-	-	-	-	NA
Synaptophysin	-	-	-	-	-	-	-	NA
TTF1	-	-	-	-	-	-	-	NA
Uroplakin-2	NA	NA	NA	NA	NA	NA	NA	-
WT1	+	+	+	+	+	+	+	NA

NA: Not available; -: negative; +: positive.

Discussion

This study represents the largest and most comprehensively described series of paratesticular tumors in individuals with Proteus syndrome, demonstrating that these tumors may be a more frequent and clinically significant manifestation of the disorder than previously recognized. We found that nearly one in five males (19%) in our natural history cohort had a history of surgery for a paratesticular mass, and that these tumors encompassed a histologically diverse group of neoplasms, including papillary cystadenomas, low grade papillary adenocarcinomas, and other rare entities such as adenocarcinoma with spindle cell transformation and Brenner tumors.

The detection of the somatic AKT1 c.49G>A, p.(Glu17Lys) variant in all sequenced tumors supports a direct pathogenic link between mosaic AKT1 activation and paratesticular tumorigenesis. Importantly, no secondary driver variants were identified, suggesting that AKT1 activation alone may be sufficient to initiate tumor development in these tissues. This finding aligns with prior observations in other AKT1-related overgrowths in Proteus syndrome, where hyperactivation of the PI3K-AKT pathway drives tissue overgrowth in the absence of additional oncogenic drivers.²

The diagnosis of paratesticular epithelial tumors, particularly in the context of a rare disorder such as Proteus syndrome, presents a significant diagnostic challenge. The differential diagnosis for epithelial neoplasms in the paratesticular region includes benign lesions such as epididymal cystadenomas, serous cystadenomas, and rete testis adenomas, as well as malignant entities such as adenocarcinomas of the rete testis, malignant mesotheliomas, and metastatic carcinomas from intra-abdominal or retroperitoneal primaries.¹² Distinguishing among these lesions can be particularly difficult when tissue architecture is distorted by chronic overgrowth or previous surgery, as is often the case in Proteus syndrome. Immunohistochemistry can aid in the differentiation, with markers including PAX8, CK7, calretinin, and WT1 helping to distinguish Müllerian-type tumors from mesothelial or metastatic lesions.

The paratesticular tumors in individuals with Proteus syndrome demonstrate overlapping features with serous borderline tumors (SBT) of Müllerian origin, but with key morphologic distinctions that inform the differential diagnosis.^{12, 13} Both tumor types exhibit papillary architecture, psammomatous calcifications, and a Müllerian immunophenotype (PAX8+, WT1+, ER/PR+). However, a characteristic feature of ovarian and paratesticular SBT is the presence of intra-cystic hierarchical papillary branching (where larger papillae progressively branch into medium and smaller papillae, with epithelial stratification and detached small epithelial clusters or individual cells along the lining) and/or intra-cystic micropapillary architecture, characterized by elongated micropapillary chains radiating from large papillae (the “Medusa head” appearance) and sheets of micropapillae situated between these structures. These characteristic types of architecture were not observed in any of the tumors from our Proteus syndrome cohort. Instead, the tumors in Proteus individuals displayed two distinct histologic profiles: (1) papillary cystadenomas with exophytic papillary growth limited to the tunica albuginea or cyst lining, and (2) adenocarcinomas with either solid growth and desmoplastic stroma or, in one case, spindle cell transformation. Furthermore, the qualitative degree of intra-cystic papillary architecture in the Proteus syndrome-associated adenocarcinomas was beyond that usually seen in SBTs. Even tumors with complex papillary architecture and mild to moderate cytologic atypia in our series lacked the epithelial-stromal interface typical of borderline tumors.

While none of the tumors originated in the epididymis, the papillary cystadenomas of our series share certain overlapping features with papillary cystadenomas of the epididymis (PCE), a rare benign neoplasm associated with von Hippel-Lindau (VHL) disease.^{14, 15} Both tumor types have intracystic papillary projections and occur in the paratesticular region. However, PCE is composed of papillae lined by clear cells, with morphologic resemblance to clear cell renal cell carcinoma and clear cell papillary renal cell tumor.¹⁵ While clear cytoplasm was a shared feature, the tumors from individuals with Proteus syndrome exhibited areas with eosinophilic cytoplasm, nuclear stratification, variable nuclear atypia and ciliated apical surfaces. By immunohistochemistry, both tumor types share diffuse labeling for PAX8 but PCE are negative for WT1 and diffusely positive for CK7 and CAIX, while tumors in Proteus syndrome were consistently diffusely positive for WT1, frequently positive for ER/PR, and only focally/patchy positive for CK7 and CAIX.

Although none of the tumors involved the rete testis, another potential differential diagnosis is with adenocarcinoma of the rete testis (ACRT), particularly those with glandular or papillary features and solid architectural components. However, there are several morphologic and immunohistochemical distinctions that help differentiate ACRT from tumors in Proteus syndrome. In the series by Chen et al., ACRT demonstrated a broad morphologic spectrum, including mucinous, and carcinoid-like features, with the presence of signet-ring cells, mucin pools, and transitional zones from benign rete epithelium, which were not observed in any tumors from those in Proteus syndrome.¹⁶ Additionally, ACRT is more common in and typically a disease of older individuals (median age 54 years), associated with aggressive clinical behavior. Importantly, ACTR can show frequent expression of WT1 and PAX8 limiting the utility of IHC in this differential diagnosis.¹⁷

Sertoliform papillary cystadenoma of the rete testis, a rare benign lesion, also enters the differential diagnosis. Morphologically, these lesions often display tubular and papillary structures with a pseudostratified columnar epithelium resembling Sertoli cells, frequently with eosinophilic cytoplasm and round to oval nuclei.¹⁸ They may be mistaken for low-grade malignancies due to their complex architecture, but typically lack the cytologic atypia, mitoses, or stromal invasion. While some Sertoliform cystadenomas may show clear cytoplasm or cribriform pattern, they are generally distinguished by their location in the rete testis and often lack WT1 expression and have variable expression of PAX8, and express inhibin and calretinin, while the tumors in Proteus syndrome had the opposite immunophenotype.¹⁹

Another important differential diagnosis is paratesticular mesothelioma. Both tumor types could share several morphologic and immunophenotypic features. Morphologically, both are characterized by papillary architecture, and both could show solid growth and reactive stroma. Immunophenotypically, both are positive for WT1 and could express PAX8. However, calretinin, a marker that is frequently positive in mesotheliomas, was uniformly negative in Proteus syndrome associated tumors.

Another differential diagnosis to be considered is with sex cord stromal tumors. While they share expression of several common immunohistochemistry markers including WT1 and SF1, a main difference is that all the paratesticular tumors in Proteus syndrome presented in the paratesticular tissue, and none involved the testicular parenchyma. In contrast, sex cord stromal tumors occur, in the great majority of cases, within the testis. Additionally, sex cord stromal tumor does not typically display the papillary architecture seen in Proteus syndrome associated paratesticular tumors.^{20, 21}

In our series, careful pathologic review and molecular confirmation of AKT1 variant mosaicism supported a primary origin and syndromic association, but such tumors could easily be misclassified in the absence of clinical context. Given the histologic overlap with ovarian and Müllerian-type tumors, including borderline and Brenner tumors, pathologists and clinicians should maintain a high index of suspicion for underlying genetic syndromes when evaluating unusual paratesticular neoplasms, particularly in children or those with other features of overgrowth; although paratesticular tumor was not the initial clinical presentation in any of the patients in this series.

The clinical behavior of these tumors was indolent overall, with no deaths attributable to paratesticular neoplasms and a long median interval to recurrence or second tumor (7.1 years). However, the recurrence rate of 50% underscores the importance of long-term surveillance, especially given the possibility of histologic progression or transformation in subsequent lesions. Notably, two individuals developed recurrent tumors that were clinically indolent but with a carcinoma histology, raising the possibility of a neoplastic continuum in some cases, a feature that we observed and described in a subset of the tumors reported here.

Our findings also highlight the challenge of distinguishing true neoplasia from overgrowth in the setting of Proteus syndrome. The small tumor size (median 1.6 cm), unilateral presentation, and slow growth observed in this series suggest that paratesticular tumors in Proteus syndrome may often be managed conservatively, particularly when radiologic and clinical features do not indicate aggressive behavior. However, in some cases, surgical resection may be warranted for symptomatic lesions or those increasing in size. The presence of variable histology in some of the tumors presented in this series suggests that partial removal risks the possibility of leaving behind a higher grade or malignant component. The lack of overt vascular invasion and diffuse stromal invasion supports a surgical approach trying to preserve the testis if not grossly involved. Additionally, tumors were managed with surgery alone, except in a single individual who received adjuvant chemotherapy where the benefit was uncertain.

In conclusion, paratesticular tumors represent a relatively common but under-recognized manifestation of Proteus syndrome. These tumors are frequently associated with somatic AKT1 variants and can exhibit a wide range of histologic patterns and clinical behaviors. Given the potential for recurrence, long-term monitoring is warranted. Future studies with larger, multi-institutional cohorts are needed to refine management strategies and elucidate the natural history of these tumors in the context of mosaic AKT1-related overgrowth.