Cancer in Multilineage Mosaic RASopathies due to Pathogenic Variants in HRAS or KRAS: A Systematic Review and Meta-analysis

Germline pathogenic variants (PV) in various RAS–MAPK pathway genes lead to a growing group of developmental disorders collectively called “RASopathies” (1). For example, HRAS germline PVs cause Costello syndrome (CS), and KRAS PVs cause Noonan and cardiofaciocutaneous syndromes (1).

Mosaic RASopathies are conditions caused by postzygotic PVs in RAS–MAPK pathway genes arising during embryonic development (2). In contrast to germline RASopathies, the spectrum of PVs typically found in mosaic RASopathies overlaps with the spectrum of somatic mutations in cancer, most of which are considered incompatible with prenatal survival when they are present in the germline. If the mosaic PV in a RAS–MAPK pathway gene occurs during organogenesis, it can lead to organ-specific manifestations such as arteriovenous malformations or congenital pulmonary airway malformations (3, 4). In contrast, PVs arising before organogenesis can affect multiple cell lineages and have broader clinical manifestations (5). Depending on the clinical presentation and involved organs, these multisystem conditions, most of which were described clinically before underlying mechanisms were first identified, have various names and include cutaneous skeletal hypophosphatemia syndrome (epidermal nevi, melanocytic nevi, or hypophosphatemic rickets), encephalocraniocutaneous lipomatosis (anomalies involving the eyes, skin, and central nervous system), epidermal nevus syndrome (epidermal nevi and involvement of the eyes, bones, urogenital tract, and central nervous system), keratinocytic epidermal nevus syndrome (keratinocytic nevi and central nervous system, eye, and bone involvement), mosaic CS (mental disability, facial features, heart defects, and skin and musculoskeletal abnormalities), oculoectodermal syndrome (epibulbar dermoids or aplasia cutis congenita), phacomatosis pigmentokeratotica (speckled lentiginous nevus and nevus sebaceous and brain, eye, and skeletal abnormalities), and Schimmelpenning–Feuerstein–Mims syndrome (anomalies involving the skin, brain, eyes, and bones; ref. 2). Here, we analyzed the cancer risk and spectrum in patients with multilineage mosaic RASopathies due to PVs in HRAS or KRAS using a pooled analysis.

Details of the search process are provided in Fig. 1. We included all identified articles describing detailed clinical and genetic information on individuals with a mosaic RASopathy due to somatic mosaic PVs in HRAS or KRAS. We included patients with documented clinical involvement in at least two organ systems and/or at least moderate skin involvement, as defined by Martínez-Glez and colleagues (6), because early mosaicism involving multiple lineages can be inferred. Cohort studies, case reports, and case series published between January 1, 1971, and December 31, 2023, were included. In several instances, mosaic PVs in HRAS or KRAS were found in patients who were already described in the literature several years previously. Where appropriate, we used the original publication to obtain clinical details on individual patients. The PubMed search was conducted between October 2023 and January 2024. The following search terms were used: mosaic rasopath*, Schimmelpenning-Feuerstein-Mims syndrome, cutaneous skeletal hypophosphatemia syndrome, phacomatosis pigmentokeratotica, encephalocraniocutaneous lipomatosis, oculoectodermal syndrome, epidermal nevus syndrome, mosaic CS, nevus sebaceous syndrome, keratinocytic epidermal nevus syndrome, and encephalocraniocutaneous syndrome. A description of these conditions can be found in a recent review article (2). The following information was collected: PubMed indentifier number (PMID), sex, gene and substitution, syndrome name, organs involved by mosaic, first cancer type (ICD-10), location of first cancer, age at diagnosis of first cancer, second cancer type (ICD-10), age at diagnosis of second cancer, age at death, age at last follow-up, and publication date.

All identified cases were used for qualitative pooled analysis. Cancer-specific hazard rates and cumulative incidence curves were calculated as described previously (7). We calculated the standardized incidence rate (SIR) comparing the incidence of first cancer in cases (only) in the literature with the expected incidence of cancer in the Surveillance, Epidemiology, and End Results Program database using SAS (version 9.4) as recently described (8). For this analysis, birth dates were inferred based on the details in the publication.

Raw data are available in the Supplementary Table S1 and upon request to the corresponding author.

Our search identified 53 articles describing 69 patients with a multilineage mosaic RASopathy due to a PV in HRAS or KRAS (Fig. 1; refs. 9–61). Details on each patient are provided in Supplementary Table S1. Demographic characteristics are summarized in Table 1. A total of 12 primary cancer and two second cancer cases were observed. The most common primary cancer type was rhabdomyosarcoma (RMS), which was diagnosed in seven patients. The mean age at first RMS diagnosis was 13 months (range 1–48 months). One patient had a second RMS at 12 years of age. Notably, in all seven patients, RMS occurred in the urogenital region (Supplementary Table S1). Nonmelanoma skin cancer was diagnosed in three patients. Additional cancer types included Wilms tumor (n = 1) and urothelial carcinoma (n = 1, the same patient had a second urothelial carcinoma later in life). SIRs are given in Table 2. The highest SIR value (800; 95% confidence interval, 320–1648) was observed for RMS. PVs in HRAS and KRAS genes reported in 69 multilineage RASopathy literature cases are shown in Fig. 2. The HRAS p.Gly13Arg substitution was the most common HRAS alteration among HRAS PV–positive multilineage RASopathy literature cases with or without cancer. Among KRAS PV–positive multilineage RASopathy literature cases with and without cancer, the KRAS p.Gly12Asp substitution was the most common alteration. As shown in Fig. 3, hazard rates peaked during the early first decade of life. By 20 years of age, 20% of patients (95% confidence interval, 4%–37%) had developed a malignancy. Most cancers occurred in the early first decade, mainly driven by RMS in young children. Subsequently, the cumulative cancer incidence increased at a slower pace, mainly driven by urothelial and skin cancers.

We conducted a systemic review to assemble a retrospective pooled cohort of patients with HRAS or KRAS PV–positive mosaic RASopathies. Our analysis suggests that these patients have a strongly increased RMS risk predominantly during the early first decade of life. Given the known role of somatic RAS mutations in childhood RMS, this association is highly plausible (62–64). Somatic RAS–MAPK mutations are present in approximately 30% of patients with fusion-negative RMS, and somatic mutations in KRAS or HRAS are particularly common in infantile RMS (64). HRAS is located on chromosome 11p15, a region commonly characterized by paternal uniparental disomy preceeding the acquisition of somatic HRAS mutations in sporadic RMS (63). Notably, all RMS affected the urogenital region. The reason for the clustering of RMS in this specific anatomic location is unknown. Sporadic RMS can arise in any anatomic location; however, the genitourinary region is commonly involved in children with embryonal RMS (64). Later in life, our findings suggest an increased risk of nonmelanoma skin cancer; however, given the young median age of our cohort, our findings may represent an underestimate. We cannot rule out that exposure to genotoxic treatment further increases the cancer risk of subsequent neoplasms in cancer survivors with mosaic RASopathies as has been described for the RASopathy Neurofibromatosis type 1 (65).

We noted genotype–phenotype correlations. Most cancers were diagnosed among carriers of mosaic PVs predicting HRAS p.Gly13Arg or KRAS p.Gly12Asp substitutions. Notably, there were several patients with a mosaic KRAS PV predicting a biochemically distinct KRAS p.Ala146Thr substitution (66), which seems to be selected among patients with certain types of mosaic RASopathies (27). No cancers were reported among patients with this mosaic PV.

Cancer surveillance guidelines for RASopathies are available (67, 68), but mosaic RASopathies are distinct. To date, the cancer risk in patients with multilineage mosaic RASopathies due to PVs in HRAS and KRAS has not been quantified formally. Morren and colleagues recently conducted a literature review of cases of syndromic epidermal nevi and related disorders and noted two RMS and three urothelial carcinomas among 28 patients with a mosaic PV in HRAS and three RMS, two Wilms tumors, and one urothelial carcinoma among 38 patients with a mosaic PV in KRAS. These authors recommend cancer surveillance, especially in patients with keratinocytic epidermal nevi involving at least three body regions, or face and neck sebaceous nevus, in patients with congenital pigment lesions or vascular malformations or oculoectodermal syndrome and encephalocraniocutaneous lipomatosis (69).

Based on our analysis, we propose that patients with KRAS or HRAS PV–positive mosaic RASopathies should be offered cancer surveillance. We recommend offering cancer surveillance to any patient with a documented mosaic PV in HRAS or KRAS with clinical involvement of at least two cell lineages or organ systems and/or at least moderate skin involvement, as defined by Martínez-Glez and colleagues (6).

The surveillance protocol should start shortly after birth or as soon as the diagnosis is established. In order to detect RMS of the urogenital region, we recommend sonography of the pelvis every 3 months (quarterly) until the fifth birthday. This is in analogy to the RMS surveillance recommended for children with CS (67, 68). Regional MRI should be offered if sonography is nondiagnostic, or in case of symptoms. We also recommend annual clinical exams. Due to the possible elevated skin cancer risk, these exams should include evaluation of the skin and be performed by dermatologists. The dermatologist should determine the frequency of necessary follow-up based on the individual skin manifestation and type of involvement. In addition, we recommend following the CS bladder cancer screening guidelines for patients with multilineage mosaic HRAS PVs (annual urinalysis with cytology starting at 10 years of age; ref. 68).

Our findings have several limitations: (i) Publication bias may have led to an overestimation of the cancer risk. Nevertheless, the fact that we observed a clear cancer pattern in this retrospective cohort, especially urogenital RMS occurring early in life, suggests that the estimated risk magnitude is real. However, prospective confirmatory studies are required. (ii) Despite our best efforts, we cannot rule out that publications were missed. In addition, our analysis focused on PubMed, and other databases were not searched. (iii) We included patients who differed with regard to their organ involvement. Thus, the group is heterogeneous. (iv) The analysis has intrinsic limits to precision of estimates because the disease is so rare. (v) Tumor specimens were not available. Tissue blocks from affected patients would allow us to look for mechanistic evidence of a causative relationship between the mosaic PV(s) and subsequent tumorigenesis. Despite these limitations, the data presented provide compelling evidence that patients with multilineage KRAS or HRAS PV–positive RASopathies are at significantly increased cancer risk, particularly RMS of the urogenital region during the first 4 years in life and skin cancer during adulthood. Ultimately, cancer surveillance is indicated.

D.R. Stewart performs contract work as a clinical geneticist for Genome Medical, a telehealth company. It is a national clinical genetics practice, not a drug, device, or trial company. This relationship is reviewed every year by internal ethics processes. No disclosures were reported by the other authors.

J. Windrich: Investigation, writing–review and editing. G.M. Ney: Formal analysis, investigation, writing–review and editing. P.S. Rosenberg: Formal analysis, writing–review and editing. J. Kim: Formal analysis, writing–review and editing. M. Zenker: Writing–review and editing. D.R. Stewart: Supervision, writing–review and editing. C.P. Kratz: Conceptualization, supervision, funding acquisition, investigation, writing–original draft.

C.P. Kratz has been supported by the Deutsche Kinderkrebsstiftung (DKS2024.03). This work was supported by the Intramural Research program of the Division of Cancer Epidemiology and Genetics, NCI, Rockville, Maryland.