Deep Phenotyping of Pathology-Confirmed Benign Lesions in PTEN Hamartoma Tumor Syndrome Patients.

Ane J. Schei‐Andersen: conceptualization, data curation, methodology, formal analysis, investigation, writing – original draft, review and editing, visualization. Janneke H. M. Schuurs‐Hoeijmakers: writing – review and editing, supervision. Rachel van der Post: writing – review and editing. Arjen R. Mensenkamp: writing – review and editing. Jolanda Schieving: writing – review and editing. Nicoline Hoogerbrugge: conceptualization, writing – review and editing, supervision. Janet R. Vos: conceptualization, data curation, writing – review and editing, supervision. PTEN Study Group: review and editing. The work reported in the paper has been performed by the authors unless clearly specified in the text.

This study (CMO light, dossier number 2019‐5291) was approved by the Medical Ethics Review Committee Oost‐Nederland, based on the criterium that patients consent to (opt‐in) or did not object to (opt‐out) secondary use of health data. The study was also reviewed by the scientific council and privacy committee of Palga.

This retrospective cohort study consisted of 379 patients tested for PTEN between 1997 and 2020 with a confirmed likely pathogenic/pathogenic (LP/P) PTEN germline variant. Patients were tested either due to personal medical history or cascade testing of the familial pathogenic PTEN variant. Variants were classified according to ACMG guidelines or the ClinGen PTEN Variant Interpretation Guidelines for PTEN, dependent on the time of diagnosis. The genetic testing results were collected from the diagnostic laboratory (386 patients with a LP/P variant in PTEN) and pathology reports were collected from the Dutch nationwide pathology databank (Palga) (381/386 patients). Patients with no testing result, a variant of uncertain significance (VUS), pathogenic variants not in PTEN , or no available pathology reports containing benign lesions (3 patients) were excluded. In total, 379 of 386 patients (98%) with a LP/P PTEN variant were included. An overview of the cohort selection is presented in Figure  1 . Cohort selection overview. Cohort selection of included patients. DNA testing was performed in the genome diagnostics laboratory of the Radboudumc between 1997 and 2020. 379 (98%) of 386 patients were included in the study. Excluded patients were excluded based on pathology database linkage and/or pathology report availability. Excerpt: Pathology excerpt (description of findings) of tissue, either through biopsy or resection; LP/P: Likely pathogenic/pathogenic; Palga: The Dutch nationwide pathology databank; PHTS: PTEN Hamartoma Tumor Syndrome. The pathology‐confirmed benign lesions were divided into PHTS‐related lesions based on the current Dutch PHTS diagnostic guideline (Table  1 ) and other observed lesions. The other observed lesions were classified by ICD‐10 codes based on the available information in the pathology reports. The age of the excerpt was based on the creation date of the abstract from the pathology report from which we calculated the patients' age. This was used in place of the age of diagnosis as there may be discrepancies between initial patient contact and pathological diagnosis. Patients were stratified by sex and index status. The last year of follow‐up was defined as the year of the pathology report request (2020). Age at last follow‐up was defined as the year of pathology report request minus year of birth. Selection bias was addressed by separating index cases (i.e., first person in a family to undergo genetic testing) and non‐index cases. PHTS clinical criteria of the Dutch diagnostic guideline. Data was analyzed in R Studio. Differences in continuous data were assessed by the Wilcoxon rank‐sum test and categorical data by Chi‐Squared or Fisher's exact test where appropriate. The Kruskal–Wallis test was used to assess frequencies stratified by sex and index status. A Bonferroni corrected p ‐value of 0.0002 was considered statistically significant.

Of the 379 included PHTS patients, 55% were female and 45% were male. The median age at last follow‐up overall was 38 years (IQR 18–53), 42 years (IQR 23–56) for females and 25 years (IQR 14–49) for males ( p  < 0.0002). Approximately half (56%) of the patients were index cases, 29% were female and 27% were male. Males were significantly younger at genetic testing compared to females (14 years (IQR 5–40) vs. 34 years (IQR 15–45), p  < 0.0002) although testing was performed in childhood and in adulthood in both sexes. Index cases and index males were significantly younger at genetic testing as well as last follow‐up compared to non‐index cases and index females ( p  < 0.0002) (Table  2 ). Patient characteristics. Note: Bold values indicate significant P ‐value after bonferroni correction. The most frequently observed PHTS‐related benign lesions were gastrointestinal lesions (54%) followed by skin (47%), thyroid (31%) and vascular lesions (26%). Approximately 57% of females developed benign breast lesions. Frequencies of PHTS‐related lesions are presented in Table  3 and Supporting Information Table  1 . The frequencies of PHTS‐related lesions in this cohort, PHTS patients from literature, and the general population are presented in Table  4 . There were no significant statistical differences observed between sexes or by index status in any of the PHTS‐related lesions. Frequency of patients per benign lesion and excerpt age. Overview of the frequency of benign lesions in the current cohort, PHTS patients from literature and the general population. Note: Listed are the estimates for PHTS related lesions and for uterine lesions based on frequencies observed in this cohort, PHTS estimates from literature and estimated general population prevalences. Abbreviation: N/A: not available. Only a single study with an available estimate. Gastrointestinal lesions were observed in 54% of the total cohort and in 80% of patients above surveillance age (> 40 years): 85% of females and 72% of males. The most frequent group of polyps were hamartomatous polyps (25%, 43 years (IQR 35–50)), consisting of hamartomatous polyps NOS (9%), ganglioneuromas (9%) and inflammatory/juvenile polyps (7%). Lymphoid polyps were present in one patient. Adenomas were observed in 15% (46 years (IQR 38–56)), followed by hyperplastic polyps (9%, 48 years (IQR 45–56)), sessile serrated lesions (3%, 49 years (IQR 46–57)) and glycogenic acanthosis (1.6%, 42 years (IQR 35–52)). The number of polyps per patient per pathology excerpt ranged from 1 to 37 polyps and 1 to 11 for gastrointestinal adenomas. All gastrointestinal lesions were slightly more frequent in females, except glycogenic acanthosis, which was slightly more frequent in males compared to females (3% vs. 0.5%). Benign skin lesions were present in 47% of the PHTS patients and consisted of fibromas (20%, 33 years (IQR 22–42)), lipomas (15%, 24 years (IQR 10–41)), papillomas (7%, 30 years, (IQR 24–38)), trichilemmomas (3%, 35 years (IQR 30–37)), skin tags (2%, 30 years (IQR 19–37)) and one patient with acanthosis nigricans (11 years). Fibromas were more frequent in females compared to males (23% vs. 15%) although males were younger at fibroma excerpt (26 years (IQR 13–34) vs. 36 years (IQR 27–42)). This remained when stratifying for index status (Figure  2 ). Males were also younger at lipoma excerpt compared to females (9 years (IQR 5–34) vs. 30 years (IQR 22–41)) also when stratifying for index status (Figure  3 ). Females were younger at trichilemmoma excerpt (31 years (IQR 28–36) vs. 38 years (IQR 38–49)). Non‐index cases were younger at papilloma excerpt compared to index cases (14 years (IQR 8–26) vs. 35 years (28–40)). Fibroma age groups by sex and index status. The blue bar indicates number of patients per age group, stratified by sex (horizontally) and index status (vertically) with a pathology‐confirmed fibroma. A total number of 74 patients had one or more pathology‐confirmed fibromas. Reflective of the results in Table  3 , fibromas are more frequent in females and especially in index females. Lipoma age groups by sex and index status. The blue bar indicates number of patients per age group, stratified by sex (horizontally) and index status (vertically) with a pathology‐confirmed lipoma. A total of 57 patients had one or more pathology‐confirmed lipomas. Reflective of the results in Table  3 , lipomas are more frequent in females although males were younger at lipoma biopsy/removal. Thyroid lesions were observed in 31% of the total cohort and in 35% of patients above surveillance age (> 11 years): 42% of females and 24% of males. Thyroid lesions consisted of benign multinodular hyperplasia (20%) with a median excerpt age of 38 years (IQR 23–49), follicular adenomas (10%) at a median excerpt age of 35 years (IQR 23–44) and Hashimoto's thyroiditis (1.3%) with a median excerpt age of 41 years (IQR 40–42). Vascular malformations were present in 26% of patients, consisting of arteriovenous malformation, lymphangiomas, and hemangiomas at a median age of 27 years (IQR 20–35) and were more frequent in females compared to males (34% vs. 16%) although not statistically significant. Eight patients (2%, six females) presented with Lhermitte‐Duclos disease at a median age of 36 years (IQR 34–41). Benign breast lesions were observed in 57% of the females and in 76% of females above surveillance age (> 25 years) in this cohort, most frequently as fibrocystic breast disease (31%) at a median age of 38 years (IQR 31–45), fibromas/fibroadenomas (19%) at a median age of 28 years (IQR 19–39) and intraductal papillomas (6%) at a median age of 27 years (IQR 20–39). Non‐index cases trended toward being younger compared to index cases at fibroma/fibroadenoma excerpt (23 years (IQR 17–36) vs. 31 years (IQR 21–43)). No benign breast lesions were observed in males. In addition to the PHTS‐related lesions, female‐specific benign lesions were also observed with endometrial hyperplasia (16%) at a median age of 46 years (IQR 38–51), uterine polyps (13%), uterine fibroids (7%) at a median age of 41 years (IQR 35–49), endometriosis (5%) and ovarian cysts (5%), cervical polyps (2%) and benign neoplasms of the ovary (3%) being most frequent. Index cases trended toward being younger at uterine fibroid excerpt (41 years (IQR 37–49) vs. 52 years (IQR 46–64)). Other non‐PHTS related benign lesions were observed in 32% of patients. The most frequently observed lesions were epidermal cysts (7%), seborrheic keratosis (3%) and ganglion (1%). No statistically significant differences were observed between sexes or by index status. The frequencies of other observed lesions are presented in Supporting Information Table  2 .

This study consisted of a large cohort of 379 PHTS patients with confirmed likely pathogenic/pathogenic germline PTEN variants. The data in this study was extracted from pathology reports from the Dutch nationwide pathology databank, which provides certainty of pathological diagnosis. Stratifying patients by index status and sex allowed us to correct for selection bias and assess whether lesions occur more frequently in one sex. As we had all excerpts for each patient available, we were able to collect and assess frequencies of any other occurring lesions not yet associated with the PHTS phenotype.

This large retrospective cohort study, consisting of 379 PHTS patients, assessed the spectrum of PHTS‐related and other observed pathology‐confirmed benign lesion frequencies and excerpt age stratified by sex and index status. The most frequent PHTS‐related lesions were gastrointestinal lesions (53%) followed by skin (47%), thyroid (35%) and vascular lesions (26%). Female‐specific lesions consisted mainly of fibrocystic breast disease (31%), breast fibroadenomas (19%) and intraductal papillomas (6%) were most frequent, followed by endometrial hyperplasia (16%) and uterine fibroids (7%). Of other benign lesions, uterine polyps (13%), epidermal cysts (7%), endometriosis (5%), ovarian cysts (5%), benign neoplasms of the ovary (3%), seborrheic keratosis (3%) and ganglion (1%) were the most frequent. The spectrum of observed benign lesions in this study is in line with previous studies and current PHTS diagnostic guidelines and consists of hamartomas involving several body systems and tissue types, reflective of the hamartomatous nature of this syndrome [ 2 , 6 , 10 ]. There were no major significant differences observed between females and males or between index cases and non‐index cases, indicating that the observed lesions seemingly do not have a sex preference and are not due to selection bias. Gastrointestinal lesions were observed in 54% of the patients, which is higher than that observed by Hendricks et al. (16%) and Starink et al. (43%) but much less than that observed by Bubien et al. (85%), Drissen et al. (89%) and Pilarski (95%) [ 1 , 10 , 12 , 13 , 14 ]. The discrepancies observed between the different studies may be reflective of different data types, country guidelines, and/or age of the patients. Hamartomatous polyps (25%) were most frequent, followed by adenomas (15%) which highlights the relevance of hamartomatous polyps as a major criterion of PHTS [ 1 , 6 , 10 , 13 ]. Polyps in our cohort were detected mostly above age 40, as many lesions will have been identified upon colon cancer surveillance, which starts at age 40 in the Netherlands [ 6 ]. Approximately half of the patients (47%) presented with one or more skin lesions, which is similar to that observed by Hendricks et al. (38%) and Drissen et al. (14%–78%) but less than that observed by Bubien et al. (98%) and in the review by Pilarski (≤ 100%) [ 1 , 2 , 10 , 12 ]. This is partly due to certain lesions, such as oral papillomas, acral keratosis, keratotic pits, trichilemmomas, and acanthosis nigricans, being less likely to get biopsied and are consequently less present or not present in the dataset [ 15 , 16 ]. Skin lesions overall were more frequent in females compared to males (55% vs. 38%). A probable explanation for this difference could be that females were significantly older than the males in this cohort and had more time to develop these lesions. Additionally, females may be more likely to have skin lesions removed for esthetic reasons, as well as some skin lesions (trichilemmomas and fibromas) being more frequent in females in general [ 17 , 18 ]. About 15% of the patients had one or more lipomas, which is less than that observed by Drissen et al. (39%), Bubien et al. (48%), Lim and Ngeow (34.6%–56.7%) and Starink et al. (31%) [ 2 , 10 , 13 , 19 ]. However, this is still remarkably higher than the incidence of lipomas in the general population, where it is estimated to be present in 1 in 1000 [ 20 ]. The patients developed lipomas at a young age (24 (IQR 10–41)) and males were younger at lipoma onset compared to females (9 vs. 30) as well as when stratifying for index status. Moreover, this early onset is further highlighted considering lipomas generally develop between 40 and 60 years of age in the general population [ 20 , 21 ]. Early‐onset lipomas (< 25 years) may be an indication to consider diagnostic PTEN testing. Approximately 31% of the patients presented with one or more benign thyroid lesions. This is similar to that observed by Tan et al. (35.4%) and Hendricks et al. (24%), but lower than that of Bubien et al. (71%), Starink et al. (68%) and Drissen et al. (84% & 87%), respectively [ 2 , 4 , 10 , 12 , 13 , 22 ]. This is likely due to many thyroid lesions not being biopsied or excised. In the Netherlands, thyroid cancer surveillance is initiated in childhood by physical examination from diagnosis and by yearly ultrasounds from 18 years onwards, with newer studies recommending surveillance start from as young as 12 years old [ 6 , 23 ]. Vascular malformations were present in 26% of patients with a median age of 27 years (IQR 20–35). This is similar to that observed by Bubien et al. (35%) and Drissen et al. (30%) and higher than that by Hendricks et al. (12%) and Starink et al. (18%) [ 2 , 10 , 12 , 13 ]. These estimates are still significantly higher than that observed in the general population (5%) which further highlights vascular malformations as an indication of PHTS [ 24 ]. Benign breast lesions were observed in 57% of the females in this cohort, which is more than that observed by Drissen et al. (35%) [ 25 ]. Benign breast lesions in this cohort consisted of fibrocystic breast disease (31%), fibromas/fibroadenomas (19%) and intraductal papillomas (6%). Fibrocystic breast disease (31%) was less frequent in this cohort than that observed by Tan et al. (44.9%) and Starink et al. (52%) [ 4 , 13 ]. However, fibrocystic breast disease is estimated to affect 50% of women from age 30 onwards in the general population [ 26 ]. It is likely that fibrocystic breast disease in this cohort is a reflection of the frequencies in the general population and not specific for PHTS patients. Fibroadenomas were present in 19% of PHTS females in our study, which is less than that seen by Schrager et al. (32%) and that observed in the general population (25%) [ 26 , 27 ]. Intraductal papillomas were seen in 6% of PHTS females, which is less than that observed by Schrager et al. (14%) and similar to that observed in the general population (5%–10%) [ 27 ]. Benign breast lesions in this cohort were less frequent than that previously described in the literature. However, due to the nature of this data, it may not showcase the full extent of benign breast lesions or their specificity as PHTS‐related lesions. Benign female‐specific lesions were observed, consisting of endometrial hyperplasia (16%), uterine polyps (13%), uterine fibroids (7%), endometriosis (5%) and ovarian cysts (5%), cervical polyps (2%) and benign neoplasms of the ovary (3%) being most frequent. Endometrial hyperplasia is a precursor lesion of endometrial cancer and is estimated to be present in 1.5%–5% of asymptomatic females in the general population and has been observed in 28% of PHTS females in a previous study [ 28 , 29 ]. The incidence rates of hyperplasia usually reach their peak when females are in their early 50s to 60s following menopause [ 28 ]. However, the estimates of hyperplasia in the general population are uncertain as it rarely gets detected unless it causes abnormal bleeding [ 28 ]. As PHTS females have an increased risk of developing endometrial cancer by age 60 of 12%–40%, it is not surprising that hyperplasia is increased in this cohort. The age of hyperplasia in our cohort was much younger (41, IQR 35–49) than that observed in the general population, which is likely reflective of an earlier cancer onset [ 9 , 30 ]. Additionally, as initiation of endometrial cancer surveillance begins at age 40 in the Netherlands for PHTS patients, this increased rate may indicate that surveillance is effective as hyperplasia is detected before it evolves into endometrial cancer [ 6 ]. However, due to the mostly asymptomatic nature of this lesion, it is rarely investigated, and therefore the clinical utility of endometrial hyperplasia for early recognition of PHTS patients is questionable. Uterine polyps (13%) were the most frequent non‐PHTS‐related female‐specific lesion. In a previous study on endometrial cancer surveillance in PHTS, a single uterine polyp was observed [ 29 ]. Uterine polyps were observed in 1.75% of Lynch Syndrome patients undergoing endometrial cancer surveillance due to their increased lifetime endometrial cancer risk (27%–70%) [ 31 ]. Furthermore, uterine polyps were observed in 7.8% of patients in a Danish population screening study [ 32 ]. Taken together, this indicates that female PHTS patients may develop uterine polyps at an increased rate compared to the general population.

The observed lesions in this large cohort study are reflective of current PHTS guidelines and previous literature. Males developed lipomas at a very early age, highlighting the relevance of lipomas as a PHTS‐related lesion. Females developed endometrial hyperplasia at an earlier age compared to the general population and developed uterine polyps more often compared to Lynch syndrome patients as well as the general population. No significant differences were observed between sexes or index cases and non‐index cases, indicating that lesions are PHTS‐specific and not due to selection bias. The results of this study are reflective of the previously observed phenotype and current PHTS diagnostic guidelines. Early‐onset lipomas may be useful for early detection of PHTS patients, and uterine polyps should be considered for inclusion as a PHTS‐related benign lesion.

Due to the nature of the data, a complete phenotyping of patients was not possible as there was no information on macrocephaly, developmental level, behavioral phenotype, etc. Additionally, not all benign lesions will be sampled or removed, and consequently, the results are an underestimation of the actual presence. Furthermore, some lesions will only be detected upon initiation of cancer surveillance, which is reflected in the similar age of biopsy for all gastrointestinal lesions and endometrial lesions [ 6 ]. Consequently, the true age of onset is likely earlier for these lesions.

PTEN Hamartoma Tumor Syndrome (PHTS) is a rare hereditary syndrome caused by pathogenic germline variants in the phosphatase and tensin homolog ( PTEN ) gene. Estimates suggest PHTS might be as rare as 1:200 000 worldwide, although this likely is an underestimation [ 1 , 2 ]. Current PHTS diagnostic guidelines consist of a complex system with multiple major and minor clinical criteria. The guidelines are based on expert medical opinions of the observed phenotype which is characterized by multi‐systemic overgrowth, including malignancies, benign lesions, developmental delay, and autism [ 1 , 2 , 3 , 4 , 5 , 6 , 7 ]. The benign lesions associated with PHTS are most often hamartomatous, as the syndrome name denotes. Hamartomas are benign tumor‐like lesions composed of disorganized growth of mature mesenchymal or epithelial tissues indigenous to the organ involved [ 8 ]. PHTS patients also have an increased cumulative lifetime risk by age 60 of developing female breast cancer (54%–67%), endometrial cancer (12%–40%) and thyroid cancer (5%–15%) with slight increases for colorectal cancer (0.5%–10%), renal cancer (1%–11%) and melanoma (2%–12%) [ 1 , 5 , 9 ]. As a result, PHTS patients are offered cancer surveillance and/or preventive treatment for their highest cancer risks [ 2 , 6 ]. Of the observed benign lesions in PHTS, gastrointestinal hamartomatous polyps, Lhermitte‐Duclos disease (dysplastic gangliocytoma), macrocephaly (head circumference > 97th percentile), penile freckling, and mucocutaneous lesions such as trichilemmomas, acral keratosis, neuromas, and papillomas are considered major criteria in the Dutch PHTS guideline [ 6 ]. Other benign minor criteria include autism spectrum disorder, glycogenic acanthosis, lipomas, intellectual disability (IQ < 75), testicular lipomatosis, benign thyroid lesions, vascular malformations, and benign breast lesions [ 6 ]. In order to receive a clinical PHTS diagnosis, a patient needs to have three major criteria, of which one of the criteria is macrocephaly, Lhermitte‐Duclos disease, or gastrointestinal hamartomatous polyps; or any two major criteria and three minor criteria [ 3 , 6 ]. The most frequently reported benign lesions in the literature are mucocutaneous lesions, which are estimated to affect almost 100% of PHTS patients [ 1 ]. Macrocephaly is also frequently observed and was present in 93% and 83% of PHTS patients in the studies from Bubien et al. and Drissen et al., respectively [ 2 , 10 ]. A variety of gastrointestinal lesions, including adenomas and hamartomatous polyps, have been observed in over 90% of patients [ 1 , 10 ]. Benign thyroid disease is estimated to affect 30%–68% of patients, and multinodular goiter was observed in 84% of PHTS patients in Drissen et al.'s study [ 1 , 2 , 10 ]. Vascular malformations have been observed in up to 35%, consisting of arteriovenous malformations, hemangiomas, and lymphangiomas [ 1 , 5 , 10 ]. Neurological disorders, including developmental delay, intellectual disability, and autism spectrum disorder, are estimated to affect 10%–20% of PHTS patients [ 1 , 5 ]. Sex‐specific benign lesions in the current guideline include penile freckling and testicular lipomatosis, which are both relatively rare in the general population [ 1 , 5 , 6 , 11 ]. Fibrocystic breast disease, fibroadenomas, intraductal papillomas, and hamartomas have been observed in up to 47% of female PHTS patients, along with ovarian cysts (37%) and uterine fibromas (25%) [ 2 , 10 , 12 ]. The specificity of female‐specific lesions as part of PHTS is uncertain, as these lesions frequently occur in the general population [ 1 ]. Previous studies detailing benign lesions in PHTS were often based on case reports, self‐reported questionnaires, or smaller cohorts. As benign lesions are often not in need of treatment, they are not always removed or studied in detail. As a result, the age of onset and frequency of these lesions are not clear. Some benign lesions also commonly occur in the general population, which makes their clinical value in PHTS uncertain [ 1 ]. Detailed knowledge of the benign phenotype could facilitate early recognition of PHTS patients, as some of these lesions may present before cancer development. This would provide clinicians the opportunity to provide preventative treatment and cancer surveillance before cancer development. Furthermore, it provides clinicians insight to determine whether a lesion should be considered part of a PHTS diagnosis. Therefore, the aim of this study is to investigate the spectrum, frequencies, and age at excerpt of pathology‐confirmed benign lesions in PHTS patients and assess whether recognition can be improved by clinical criteria based on pathology‐confirmed benign lesions in PHTS.

The authors declare no conflicts of interest.

Data S1. Supporting Information.