Dataset Information

Head and neck cancer of unknown primary: unveiling primary tumor sites through machine learning on DNA methylation profiles.

ABSTRACT:

Background

The unknown tissue of origin in head and neck cancer of unknown primary (hnCUP) leads to invasive diagnostic procedures and unspecific and potentially inefficient treatment options for patients. The most common histologic subtype, squamous cell carcinoma, can stem from various tumor primary sites, including the oral cavity, oropharynx, larynx, head and neck skin, lungs, and esophagus. DNA methylation profiles are highly tissue-specific and have been successfully used to classify tissue origin. We therefore developed a support vector machine (SVM) classifier trained with publicly available DNA methylation profiles of commonly cervically metastasizing squamous cell carcinomas (n = 1103) in order to identify the primary tissue of origin of our own cohort of squamous cell hnCUP patient's samples (n = 28). Methylation analysis was performed with Infinium MethylationEPIC v1.0 BeadChip by Illumina.

Results

The SVM algorithm achieved the highest overall accuracy of tested classifiers, with 87%. Squamous cell hnCUP samples on DNA methylation level resembled squamous cell carcinomas commonly metastasizing into cervical lymph nodes. The most frequently predicted cancer localization was the oral cavity in 11 cases (39%), followed by the oropharynx and larynx (both 7, 25%), skin (2, 7%), and esophagus (1, 4%). These frequencies concord with the expected distribution of lymph node metastases in epidemiological studies.

Conclusions

On DNA methylation level, hnCUP is comparable to primary tumor tissue cancer types that commonly metastasize to cervical lymph nodes. Our SVM-based classifier can accurately predict these cancers' tissues of origin and could significantly reduce the invasiveness of hnCUP diagnostics and enable a more precise therapy after clinical validation.

SUBMITTER: Stark L

PROVIDER: S-EPMC10964705 | biostudies-literature | 2024 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Head and neck cancer of unknown primary: unveiling primary tumor sites through machine learning on DNA methylation profiles.

Stark Leonhard L Kasajima Atsuko A Stögbauer Fabian F Schmidl Benedikt B Rinecker Jakob J Holzmann Katharina K Färber Sarah S Pfarr Nicole N Steiger Katja K Wollenberg Barbara B Ruland Jürgen J Winter Christof C Wirth Markus M

Clinical epigenetics 20240325 1

<h4>Background</h4>The unknown tissue of origin in head and neck cancer of unknown primary (hnCUP) leads to invasive diagnostic procedures and unspecific and potentially inefficient treatment options for patients. The most common histologic subtype, squamous cell carcinoma, can stem from various tumor primary sites, including the oral cavity, oropharynx, larynx, head and neck skin, lungs, and esophagus. DNA methylation profiles are highly tissue-specific and have been successfully used to classi ...[more]

PMID: 38528631

Similar Datasets

Project description:A subset of head and neck squamous cell carcinomas present solely as metastatic disease in the neck and are of unknown primary origin (SCCUP). Most primary tumors will ultimately be identified, usually in the oropharynx. In a minority of cases, the primary site remains elusive. Here, we examine the role of ancillary testing, including mutational signature analysis (MSA), to help identify likely primary sites in such cases. Twenty-two cases of SCCUP in the neck, collected over a 10-year period, were classified by morphology and viral status; including human papillomavirus (HPV) testing by p16 immunohistochemistry (IHC) and RT-qPCR, as well as Epstein-Barr virus (EBV) testing by EBER-ISH. CD5 and c-KIT (CD117) IHC was done to evaluate for possible thymic origin in all virus-negative cases. Whole exome sequencing, followed by MSA, was used to identify UV signature mutations indicative of cutaneous origin. HPV was identified in 12 of 22 tumors (54.5%), favoring an oropharyngeal origin, and closely associated with nonkeratinizing tumor morphology (Fisher's exact test; p = 0.0002). One tumor with indeterminant morphology had discordant HPV and p16 status (p16+/HPV-). All tumors were EBV-negative. Diffuse expression of CD5 and c-KIT was identified in 1 of 10 virus-negative SCCUPs (10%), suggesting a possible ectopic thymic origin rather than a metastasis. A UV mutational signature, indicating cutaneous origin, was identified in 1 of 10 (10%) virus-negative SCCUPs. A cutaneous auricular primary emerged 3 months after treatment in this patient. Primary tumors became clinically apparent in 2 others (1 hypopharynx, 1 hypopharynx/larynx). Thus, after follow-up, 6 tumors remained unclassifiable as to the possible site of origin (27%). Most SCCUPs of the neck in our series were HPV-associated and thus likely of oropharyngeal origin. UV signature mutation analysis and additional IHC for CD5 and c-KIT for possible thymic origin may aid in further classifying virus-negative unknown primaries. Close clinical inspection of hypopharyngeal mucosa may also be helpful, as a subset of primary tumors later emerged at this site.

Project description:ImportancePatients with suspected head and neck squamous cell carcinoma of unknown primary (HNSCCUP) may undergo tonsillectomy and tongue base mucosectomy (TBM) to help identify clinicoradiologically occult primary disease. It is hypothesized that when these diagnostic specimens are analyzed, conventional histopathological (CH) techniques risk missing small primary tumors that may be hidden in the tissue blocks.ObjectiveTo establish the outcomes of a step serial sectioning (SSS) histopathological technique vs CH when analyzing diagnostic tissue specimens from TBM and tonsillectomy performed for HNSCCUP.Design, setting, and participantsThe MOSES prospective multicenter noninterventional cohort study was conducted over a 25-month period from November 2019 at secondary and tertiary care ear, nose, and throat departments in the United Kingdom and included adults with clinicoradiologically occult HNSCCUP who were undergoing TBM.InterventionConventional histopathological techniques performed on TBM and tonsillectomy specimens at participating centers, followed by SSS performed at the central laboratory.Main outcomeIdentification of cancer on central histopathological review of TBM and tonsillectomy specimens.ResultsTissue from 58 eligible patients was analyzed (median [range] age, 58 [47-82] years; 10 women [17%]), with 20 480 sections cut in the laboratory and 4096 sections directly examined by a pathologist (median [range], 64 [28-135] per patient). The overall identification rate for TBM following SSS according to study protocol was 50.0% (95% CI, 37.5%-62.5%) and by subgroups was 42.9% (95% CI, 21.4%-67.4%) when performed following a negative bilateral tonsillectomy, 46.7% (95% CI, 24.8%-69.9%) at the same time as bilateral tonsillectomy, and 57.1% (95% CI, 36.5%-75.5%) following historic tonsillectomy. Conventional histopathological techniques at central review identified 2 undiagnosed primary tumors and revised the diagnosis of 2 other cases (1 nonmalignant and another down staged). Step serial sectioning identified a single additional tumor: an ipsilateral synchronous tongue base tumor for which a contralateral tumor had been identified on CH. Multifocal disease was seen in 5 (8.6%); all were human papillomavirus-related and in the tongue base.Conclusions and relevanceIn this multicenter cohort study of patients undergoing TBM for HNSCCUP, SSS was associated with added considerable histopathological workload with minimal additional diagnostic benefit. A second opinion for conventional histological techniques may be more beneficial. Synchronous primary disease should be considered when planning diagnostic oropharyngeal surgery for these patients.

Project description:ImportancePostoperative chemoradiation is the standard of care for cancers with positive margins or extracapsular extension, but the benefit of chemotherapy is unclear for patients with other intermediate risk features.ObjectiveTo evaluate whether machine learning models could identify patients with intermediate-risk head and neck squamous cell carcinoma who would benefit from chemoradiation.Design, setting, and participantsThis cohort study included patients diagnosed with squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx, or larynx from January 1, 2004, through December 31, 2016. Patients had resected disease and underwent adjuvant radiotherapy. Analysis was performed from October 1, 2019, through September 1, 2020. Patients were selected from the National Cancer Database, a hospital-based registry that captures data from more than 70% of newly diagnosed cancers in the United States. Three machine learning survival models were trained using 80% of the cohort, with the remaining 20% used to assess model performance.ExposuresReceipt of adjuvant chemoradiation or radiation alone.Main outcomes and measuresPatients who received treatment recommended by machine learning models were compared with those who did not. Overall survival for treatment according to model recommendations was the primary outcome. Secondary outcomes included frequency of recommendation for chemotherapy and chemotherapy benefit in patients recommended for chemoradiation vs radiation alone.ResultsA total of 33 527 patients (24 189 [72%] men; 28 036 [84%] aged ≤70 years) met the inclusion criteria. Median follow-up in the validation data set was 43.2 (interquartile range, 19.8-65.5) months. DeepSurv, neural multitask logistic regression, and survival forest models recommended chemoradiation for 17 589 (52%), 15 917 (47%), and 14 912 patients (44%), respectively. Treatment according to model recommendations was associated with a survival benefit, with a hazard ratio of 0.79 (95% CI, 0.72-0.85; P < .001) for DeepSurv, 0.83 (95% CI, 0.77-0.90; P < .001) for neural multitask logistic regression, and 0.90 (95% CI, 0.83-0.98; P = .01) for random survival forest models. No survival benefit for chemotherapy was seen for patients recommended to receive radiotherapy alone.Conclusions and relevanceThese findings suggest that machine learning models may identify patients with intermediate risk who could benefit from chemoradiation. These models predicted that approximately half of such patients have no added benefit from chemotherapy.

Dataset Information

Head and neck cancer of unknown primary: unveiling primary tumor sites through machine learning on DNA methylation profiles.

Background

Results

Conclusions

Publications

Head and neck cancer of unknown primary: unveiling primary tumor sites through machine learning on DNA methylation profiles.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets