Patients and tumor samples
One hundred and thirteen patients underwent radiation therapy (RT) or CCRT at the National Cancer Center Hospital, Tokyo, between January 2008 and December 2017 (Figure S1). Seventy of the 113 Japanese patients had locally advanced cervical cancer with FIGO stage IIIA to IVA, and these patients were recruited for this study. Patients received external beam RT and brachytherapy 36, and most of the chemotherapy regimen was cisplatin-based. Clinicopathological data, including age at histological diagnosis, FIGO stage, histological subtypes, status of pelvic/para-aortic lymph nodes, tumor size, treatment, and follow-up, were obtained from the electronic medical records. Cervical tumor specimens were collected by punch biopsy of the tumor before CCRT. The specimens were fixed in 10% neutral buffered formalin and embedded in paraffin (FFPE).
Treatment regimens for CCRT and RT
All patients, except one, received both external beam RT and brachytherapy (intracavitary brachytherapy or intracavitary/interstitial brachytherapy). The initial 20–40 Gray (Gy) was delivered to the whole pelvis using the 4-field box technique, followed by a 40 mm-wide midline block until pelvic side wall dose of 50 Gy. If enlarged lymph nodes were present, an additional 6–10 Gy was delivered with smaller fields. After the initiation of the midline block, a total of 3–4 sessions of brachytherapy were performed in 1–2 sessions per week, and the dose per fraction was 6 Gy. All brachytherapy was performed by an 192Iridium remote afterloading system (RALS, MicroSelectron, HDRTM, Elekta, Veennendaal, The Netherlands). The concurrent chemotherapy regimen was usually 40 mg/m2/week of cisplatin, whereas some patients received other regimens, such as carboplatin, cisplatin plus tegafur, gimeracil, oteracil, and cisplatin plus fluorouracil.
DNA preparation and next-generation sequencing
Genomic DNA was extracted from FFPE tumor tissues using the QIAamp DNA FFPE tissue kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions. Purified genomic DNA (50 ng) obtained from tumor tissues was used for library construction using the Ion AmpliSeqTM Cancer Hotspot Panel v2 (Thermo Fisher Scientific, Waltham, MA, USA), which targets approximately 2,800 COSMIC mutational hotspot regions of 50 cancer-related genes. An Ion AmpliSeqTM Custom Panel, designed for the TP53 gene (coverage: all coding regions) using Ion AmpliSeqTM Designer (https://www.ampliseq.com), was also used. Sequencing was performed on the Ion Proton platform (Thermo Fisher Scientific). For quality control, samples with a mean read depth of coverage over 1000 and a base quality score of 20 (with ≤ 1% probability of being incorrect), which accounted for 90% of the total reads, were selected.
Locally advanced cervical cancer in TCGA database
We selected 54 cases with locally advanced cervical cancer registered in TCGA database. Somatic mutations called from whole genome sequencing and whole exome sequencing data available in TCGA database were downloaded as a mutation annotation format (MAF) file via the cBioPortal for Cancer Genomics (http://www.cbioportal.org).
Classification of oncogenic/pathogenic mutations
Data analysis was carried out using the Torrent Suite Software v5.0.4 (Thermo Fisher Scientific). We selected mutations that met the following criteria: the frequency of variant alleles was more than 4% in tumor tissues; single nucleotide polymorphisms were excluded if they showed a threshold allele frequency ≥ 0.01 in either the National Heart, Lung, and Blood Institute (NHLBI) Grand Opportunity Exome Sequencing Project (ESP6500; http://evs.gs.washington.edu/EVS/) or the integrative Japanese Genome Variation Database (iJGVD, 20181105; https://ijgvd.megabank.tohoku.ac.jp/). The variants have been registered as “pathogenic/likely pathogenic variants” in ClinVar 37 or “oncogenic/likely oncogenic variants” in OncoKB (http://oncokb.org) databases using the OncoKB annotator commit 8910b65 (accessed on June 29, 2019). All selected variants were validated using the Integrative Genomics Viewer (IGV; http://www.broadinstitute.org/igv/).
Definition of actionable mutations
OncoKB is a precision oncology knowledge database that contains information on the effects and treatment implications of specific genomic alterations in cancer patients. Somatic mutations and copy number alterations have been categorized into four evidence levels. In the present study, genetic aberrations with evidence levels 1–3B according to OncoKB level of evidence V2 were designated as actionable mutations for molecular-targeting drugs 38.
Immunohistochemical (IHC) staining of p53
IHC staining was performed on FFPE specimens. Representative whole 4 μm-thick sections were analyzed. After deparaffinization, the protein expression of p53 was evaluated using a monoclonal antibody against human p53 protein (clone DO-7, Dako, Glostrup, Denmark). IHC staining was performed using a Dako autostainer (Dako, CA, USA) and visualized using EnVision Detection System (Dako), according to the manufacturer’s instructions. The slides were counterstained with hematoxylin. Staining for p53 expression was evaluated as wild-type or mutant 27. Scattered, mosaic, mid-epithelial p53 expression was considered to represent the wild-type staining pattern. Mutant staining pattern was characterized by diffuse strong nuclear positivity in the basal and upper layers of the tumor cells, or complete absence of p53 staining with appropriate positive internal control.
Identification of HPV genotyping by Sanger sequencing
HPV genotyping was performed for 70 cases. Genomic DNA (10 ng) was amplified via polymerase chain reaction (PCR) using TaKaRa Taq DNA polymerase (Takara Bio Inc., Shiga, Japan) for two distinct HPV genomic regions. The HPV E6/E7 homologous region was amplified using the pU-1M/pU2R (HPVpU-1M: 5′-TGTCAAAAACCGTTGTGTCC-3′, and HPVpU-2R: 5′-GAGCTGTCGCTTAATTGCTC-3′) primer set, and the region containing the HPV L1 gene was amplified using the GP5+/GP6+ (GP5+: 5′-TTTGTTACTGTGGTAGATACTAC-3′, and GP6+: 5′-GAAAAATAAACTGTAAATCATATTC-3′) primer set. PCR reactions were performed using the TaKaRa PCR Human Papillomavirus Typing Set (TakaRa Bio Inc.). PCR products were purified using the NucleoSpin Gel (Takara Bio Inc.) or PCR Clean-up kit (Takara Bio Inc.). Sanger sequencing was performed using an ABI 3130xl DNA Sequencer (Applied Biosystems, Foster City, California, USA), according to the manufacturer’s instructions. Similarity between the obtained sequences and various HPV genotypes in the GenBank database was determined using Basic Local Alignment Search Tool (BLAST) (https://blast.ncbi.nlm.nih.gov/Blast.cgi).
Detection of high-risk HPV types in cervical cancer tissues
To clarify the frequency of HPV-positive results in these samples, we performed in situ hybridization assay for HPV detection (HPV-ISH) using HPV-III High Risk probes (Roche Diagnostics, Mannheim, Germany), according to the manufacturer’s instructions. This assay can detect high-risk HPV genotypes, including HPV-16, 18, 31, 33, 35, 45, 52, 56, 58, and 66, in cervical cancer specimens 16.
Statistical analysis
The Kaplan-Meier method was applied to estimate survival, progression-free survival (PFS), and locoregional relapse-free survival (LRFS). Differences in outcomes were compared using the log-rank test. PFS was defined as the interval from the start of first RT to either disease progression or death. Overall survival (OS) was defined as the interval from the start of the first RT to death. LRFS was defined as the interval from the start of first RT to either locoregional disease progression or death. PFS, OS, and LRFS were determined at the last contact date for each patient. Cox regression analysis was used to assess the univariate prognostic significance of survival. Using multivariate Cox proportional-hazards models, we considered each mutation status, histological subtype, para-aortic lymph node metastasis, and tumor size. The data cut-off date was January 29, 2020. Statistical analyses were performed with EZR version 1.37 39, which is based on R and R commander.