Mutational Signature Changes in Patients With Metastatic Squamous Cell Carcinoma of the Anal Canal

Abstract Purpose We examined the concordance of genetic mutations between pretreatment tumor tissue and posttreatment circulating tumor DNA (ctDNA) in patients with metastatic squamous cell carcinoma of the anal canal (SCCA) and assessed the impact of therapy on this concordance. Methods We analyzed next-generation sequencing reports from pretreatment tumor tissue and posttreatment ctDNA in 11 patients with metastatic SCCA treated at Vanderbilt University Medical Center between 2017 and 2021. Results Among the mutations identified in posttreatment ctDNA, 34.5% were also found in pretreatment tumor tissue, while 47.6% of pretreatment tumor tissue mutations were found in posttreatment ctDNA. Four patients had preservation of potentially actionable mutations in both pretreatment tissue and posttreatment ctDNA, while 7 patients had newly identified mutations in posttreatment ctDNA that were not present in pretreatment tumor tissue. Conclusion Patients with SCCA demonstrate a high degree of temporal mutational heterogeneity. This supports the hypothesis that ctDNA can serve as a real-time tracking mechanism for solid tumors’ molecular evolution in response to therapy. Our findings highlight the potential of ctDNA in identifying emerging actionable mutations, supplementing information from tissue-based genomic assessments. Further research, ideally with larger and multi-institutional cohorts, is needed to validate our findings in this relatively rare tumor type.


Introduction
In 2023, there will be an estimated 9760 patients diagnosed with anal cancer and 1870 deaths attributable to anal cancer in the US. 1 The incidence of squamous cell carcinoma of the anal canal (SCCA) in the US increased by 2.7% per year, and anal cancer-related deaths increased by 3.1% per year, from 2001 to 2015. 2 Several factors that increase the risk of developing SCCA have been identified, including more than 10 sexual partners, receptive anal intercourse before the age of 30, chronic immunosuppressed states such as solid-organ transplant or infection with human immunodeficiency virus, active tobacco use, and being older than 50 years old. 3,4][8][9][10] Immunotherapy has also shown efficacy in patients with refractory metastatic disease and is being studied, in combination with chemotherapy for use in the first-line setting for patients with metastatic disease. 11,12Achieving disease control in patients with metastatic SCCA is a challenge, with one analysis finding that patients with metastatic SCCA treated with combination chemotherapy experience a median progression-free survival of 7 months and median overall survival of 22 months. 8irculating tumor DNA (ctDNA) has been used effectively, in conjunction with imaging, to measure therapeutic response to chemoradiation in patients with locoregional anal cancer. 13etection of HPV ctDNA is an emerging tool to detect minimal residual disease in patients with low disease burden, particularly following chemoradiation, and has been shown to be a mechanism to identify patients at high risk of recurrence in this setting. 14owever, ctDNA has not been well studied in patients with SCCA as a means to understand spatial mutational e476 The Oncologist, 2024, Vol. 29, No. 4   heterogeneity (within a tumor or across sites of disease within one patient).This heterogeneity can lead to incomplete genomic information obtained via tumor biopsy. 15,16emporal mutational heterogeneity, or the evolution of variation in the genetic composition of tumor cells over time that can occur as a response to treatment and ultimately leads to resistance to a particular therapy, represents another challenge to cancer treatment. 16A way to confront these challenges may be to serially assess a tumor's mutational profile dynamically during treatment with peripheral blood measurements of ctDNA, particularly when patients exhibit disease progression, to identify newly acquired mutations that could serve as a target for therapy and allow for personalized changes in treatment while forgoing an invasive tissue biopsy (Fig. 1).
The relatively low incidence and prevalence of SCCA pose a unique challenge to characterizing detailed tumoral genomic information in this patient population.Few analyses of the genomic profile of metastatic SCCA have been performed. 17,180][21][22][23] In this study, we sought to describe how tissue-based and ctDNA-based mutational concordance, as identified by next-generation sequencing (NGS), is impacted by chemoradiation and systemic therapy in the context of patients' unique clinicodemographic features.

Methods
We identified 11 patients with histologically confirmed metastatic SCCA treated at Vanderbilt University Medical Center between 2017 and 2021 with tissue-based TEMPUS 648 gene panel reports and ctDNA-based 105 gene panel reports available for analysis.Clinicodemographic features included in our study were: HPV status, clinical stage, the time interval between tissue biopsy and ctDNA collection, the type and amount of therapy administered during the interval between tumor tissue and ctDNA collection, and the mutations identified in each sample.Mutations were classified into predetermined categories on tissue-based TEMPUS 648 gene panel reports and ctDNA-based TEMPUS 105 gene panel reports ("Somatic-potentially actionable," "Somatic-biologically relevant," and "Variants of unknown significance").We reported variant allele frequencies (VAF) if they were included in the sample's reports.

Results
Table 1 displays by patient, the source of tissue, HPV status, clinical stage at the time of sample collection, collection Interval, interval therapy, and genes in which mutations were identified.Table 2 displays the same information but includes the specific mutations and genetic aberrations identified by NGS.Eight patients had tissue obtained from their primary tumor in the anal canal, while 3 patients had tissue obtained from sites of metastasis (lung [patient 4], pelvic lymph node [patient 5], and brain [patient 6]).Nine patients were HPV positive, as indicated by p16+ status.Four (1, 2, 8, and 10) patients with mutations in tumor tissue categorized as "somatic-potentially actionable" had the same mutations in ctDNA.In 3 (1, 2, and 10) of these 4 patients, the VAF was lower in ctDNA than in tumor tissue.Two 1,10 patients with mutations in tumor tissue categorized as "somatic-biologically relevant" had the same mutations in ctDNA.In one of these patients, the VAF was lower in ctDNA than in tumor tissue.Two patients with mutations in tumor tissue categorized as "variants of unknown significance" had the same mutations in ctDNA.In one 2 of the 3 patients, the VAF was lower in ctDNA than in tumor tissue.
Table 3 displays the mean number of mutations in tissue and ctDNA, in addition to the percentage of tumor tissue mutations and ctDNA mutations found in ctDNA and tissue, respectively.Among the 11 patients in our study, there were a total of 92 mutations found in tissue and 21 mutations found in ctDNA.The mean number of mutations in tissue was 8.4 (92 mutations/11 patients = 8.4 mutations per patient), and the mean number of mutations in ctDNA was 2.6 (29 mutations/11 patients = 2.6 mutations/patient).The percentage of mutations in ctDNA found in tissue was 34.5% (10 concordant mutations between tissue and ctDNA/29 mutations in ctDNA).Importantly, only 21 of the 92 mutations in tissue were in genes that were also analyzed in the ctDNA-based TEMPUS 105 gene panel reports.Therefore, to calculate the percentage of mutations in tissue found in ctDNA, only mutations in genes that were assessed both by the tissue-based TEMPUS 648  gene panel report and ctDNA-based TEMPUS 105 gene panel report were included.Therefore, the percentage of mutations in tissue found in ctDNA was 47.6% (10 concordant mutations between tissue and ctDNA/21 mutations in tissue).

Discussion
Our analysis outlines the impact of therapy on the mutational concordance between tumor tissue and ctDNA in patients with metastatic SCCA.Importantly, our real-world cohort of patients was similar to other cohorts described in the literature and the US, with regard to rates of HPV-positivity (9 patients; 81.8%) and mutations in PIK3CA (5 patients; 45.4%) which is one of the most commonly mutated genes in SCCA. 17,18,24The most frequently mutated "somatic-potentially actionable" and "somatic-biologically relevant" genes in tumor tissue among patients in our cohort were PIK3CA (4, 5, 9, 10, and 11) and CYLD (4, 6, and 7), respectively.Notably, one prior study suggested that mutations in TP53, PIK3CA, KMT2C, KMT2D, RB1, FAT4, NF1, CDKN2A, and CTNNB1 are driver mutations of anal cancer development. 25We identified several patients with tumor tissue containing mutations in these genes.A TP53 mutation was found in one patient's 1 tumor tissue.PIK3CA mutations were found in 5 patients' (4, 5, 8, 10, and 11) tumor tissue.A KMT2C mutation and KMT2D mutation were found in 2 separate patients' tumor tissues (11 and 8, respectively).CDKN2A mutations were found in 2 patients' (1 and 10) tumor tissue.

Effect of Therapy on Concordance and VAF
VAF is a surrogate measure of the percentage of DNA in a sample that carries mutations. 26A higher VAF, as measured by ctDNA, is a marker of tumor burden and is associated with worse survival. 27Notably, genomic alterations with a VAF of less than 1% may not actually represent DNA shed from a patient's tumor but rather clonal hematopoiesis of indeterminate potential (CHIP). 28It has been shown that VAF measured by ctDNA is lower than that of tissue among samples collected before the initiation of therapy in patients with non-small cell lung cancer. 29VAF, when measured serially in patients receiving systemic therapy, has been shown to decrease over time in patients who experience a response (determined by imaging) to treatment. 30In our analysis, we found that VAF was lower in ctDNA compared to tissue in the majority of concordant mutations (except PI3KCA in patient 8 and RAD51C and JAK2 in patient 9).However, it is unclear if this difference was related to the fact that tissue VAF is higher in tissue, generally, or whether this was representative of response to therapy.

Acquisition of Previously Absent Mutations
When patients with metastatic cancer experience disease progression, it is not uncommon for clinicians to obtain tissue from a new site of disease and assess for emerging actionable mutations with NGS.The viability of utilizing ctDNA for this purpose, as well as to understand whether certain changes in a tumor's mutational profile are indicative of response (or lack thereof) to therapy, is currently being investigated in several clinical trials in patients with melanoma and NSCLC (NCT04166487, NCT04966676, NCT04093167).Seven (1, 2, 3, 6, 8, 9, and 10) patients had newly identified mutations in ctDNA that were not present in tumor tissue.
Three patients (6, 9, and 10) had potentially actionable mutations in ctDNA that were not previously present in tumor tissue.2][33] This patient was treated with pemigatinib, an FGFR2 inhibitor, and achieved a durable, objective response. 33Patient 9 was found to have a mutation in BRCA1 (c.2029_2030delinsTT p.G677L MV; VAF = 4.8%) in ctDNA that was not previously present in tumor tissue.In a meta-analysis that examined the activity of polyadenosine diphosphate-ribose polymerase (PARP) inhibitors, which have efficacy in specific patient populations with tumors containing BRCA1/2 mutations, 24 of 43 patients with tumors containing somatic BRCA mutations treated with PARP inhibitors had a response to therapy across 8 trials. 34Patient 10 was found to have a mutation in ERBB2 (.234G>T p.Q78H MV; VAF = 28.2%) in ctDNA that was not previously present in tumor tissue.Genetic alterations in ERBB2 have been established as a biomarker for targeted anti-human epidermal growth factor receptor 2 (HER2) therapy in patients with multiple types of solid tumors, including gastrointestinal malignancies such as gastroesophageal, gastric, and colorectal cancers. 35,36our patients (2, 3, 6, and 8) had mutations with a VAF greater than 45% in ctDNA that were not previously present in tumor tissue, which may indicate they were drivers of resistance to therapy and metastasis.These mutations were in KDR (c.1379G>T p.W460L MV; VAF = 50.4%),SDHA (c.1663 + 3G>C SRV; VAF = 46.8%),HNF1A (c.1336G>A p.V446M MV; VAF = 57.4%),FGFR2 (c.1364C>T p.T455M MV; VAF = 49.7%),PBRM1 (c.1409A>G p.Y470C MV; VAF = 85.1%),DDR2 (c.749T>A p.V250E MV; VAF = 49.9%), and MYCN (c.1090C>T p.P364S MV; VAF = 48.6%).The function of the proteins these genes encode varies.According to the National Center for Biotechnology Information's Gene database, KDR encodes vascular endothelial growth factor receptor 2, SDHA encodes a major catalytic subunit of a complex in the mitochondrial respiratory chain, HNF1A encodes a transcription factor required for the expression of some liver-specific genes, FGFR2 encodes fibroblast growth receptor 2, PBRM1 encodes a subunit of a chromatin remodeling complex integral for nuclear hormone-receptor mediated transcriptional activation, DDR2 encodes a receptor subclass of the receptor tyrosine kinase family, and MYCN encodes a nuclear protein involved in the regulation of transcription. 25

at Time of Sample Collection Collection interval (days) Interval Therapy (cycles) Somatic-potentially actionable (VAF), if reported) Somatic-biologically relevant (VAF, if reported) Variants of unknown significance (VAF, if reported)
Indicates a mutation observed in one patient's tissue-based TEMPUS 648 gene panel report and ctDNA-based 105 gene panel report.*Indicates mutations observed in tissue-based TEMPUS 648 gene panel reports that were not tested for in ctDNA-based TEMPUS 105 gene panel reports.**Indicates mutations observed in ctDNA-based TEMPUS 105 gene panel reports that were not present in the same patient's tissue-based TEMPUS 648 gene panel reports.Abbreviation: VAF, variant allele frequency.

Table 3 .
Mutations in tumor tissue and ctDNA.Asterisk indicates the exclusion of mutations in genes not tested for by ctDNA-based 105 gene panel reports.