Validation of a prognostic nomogram for locally advanced oropharyngeal carcinoma (OPC) treated with intensity modulated radiation therapy (IMRT) ± systemic therapy (ST).

Paolo Bossi IRCCS Humanitas Research Hospital, Medical Oncology and Hematology Unit; Department of Biomedical Sciences, Humanitas University, Rozzano, Italy info_outline Paolo Bossi, Riccardo Gili, Stefano Calza, Lisa F. Licitra, Marta Maddalo, Almalina Bacigalupo, Liliana Belgioia, Jon Cacicedo, Nadia Facchinetti, Almudena Garcia, Marc Oliva, Pierluigi Bonomo, Giuseppe Sanguineti, Pierfrancesco Franco, Athanassios Argiris, Amanda Psyrri, Ester Orlandi

Authors Paolo Bossi IRCCS Humanitas Research Hospital, Medical Oncology and Hematology Unit; Department of Biomedical Sciences, Humanitas University, Rozzano, Italy info_outline Paolo Bossi, Riccardo Gili, Stefano Calza, Lisa F. Licitra, Marta Maddalo, Almalina Bacigalupo, Liliana Belgioia, Jon Cacicedo, Nadia Facchinetti, Almudena Garcia, Marc Oliva, Pierluigi Bonomo, Giuseppe Sanguineti, Pierfrancesco Franco, Athanassios Argiris, Amanda Psyrri, Ester Orlandi Organizations IRCCS Humanitas Research Hospital, Medical Oncology and Hematology Unit; Department of Biomedical Sciences, Humanitas University, Rozzano, Italy, UO Oncologia Medica 2, IRCCS Ospedale Policlinico San Martino, Genova, Italy, University of Brescia, Brescia, Italy, Fondazione IRCCS Istituto Nazionale dei Tumori and University of Milan, Milan, Italy, Radiation Oncology Policlinico San Martino IRCCS, Genova, Italy, Radiation Oncology Department, IRCCS Ospedale Policlinico San Martino, Genoa, Italy, Radiation Oncology Department, Cruces University Hospital, Bilbao, Spain, Radiotherapists Unit, National Center of Oncological Hadrontherapy (CNAO), Pavia, Italy, Hospital Marqués de Valdecilla, Santander, Spain, Medical Oncology Department, Catalan Institute of Oncology (ICO), Hospital Duran i Reynals, L'hospitalet De Llobregat, Spain, Ospedale Di Careggi, Firenze, Italy, Radiation Oncology, IRCCS Regina Elena National Cancer Institute, Roma, Italy, Department of Translational Medicine, University of Eastern Piedmont, Novara, Italy, Novara, Italy, Department of Medical Oncology, Thomas Jefferson University, Philadelphia, PA, University of Athens, Chaidari, Greece, Clinical Department, National Center for Oncological Hadrontherapy (CNAO), Pavia, Italy Abstract Disclosures Research Funding No funding sources reported Background: The interplay among prognosticators in OPC patients (pts) may be influenced by sociodemographic, geographic, epidemiologic and genetic features. We aimed to validate an existing prognostic nomogram (PN) (Fakhry et al 2017) in a large, independent, geographically homogeneous pt cohort. Methods: We retrospectively collected data on consecutive OPC pts treated with definitive IMRT (66-72 Gy) ± systemic therapy (ST) in 14 South-European Centres from 2007 to 2019. We analysed the role of the nomogram prognostic factors (PF) (T and N stage, age, p16 status, smoking, performance status, weight loss, anemia, education, and marital status) with respect to overall survival (OS) and progression-free survival (PFS), and we also evaluated alcohol abuse, comorbidities according to ACE-27 and the presence of a caregiver as added PF. The PN was evaluated in terms of discrimination performance (Harrel C-index) and calibration (ratio O/E and calibration line). Model was recalibrated evaluated via optimism adjusted C-index. Results: We considered 786 pts, with a median follow-up of 63 months (0-189). The table reports pts characteristics. The PN showed a good discrimination for OS and PFS both at 2- and 5-years, not dissimilar from original external validation, with a C-index of 0.75 (CI 95% 0.70-0.79,) and 0.68 (0.64-0.72), respectively. Adding alcohol abuse, comorbidities and caregiver presence did not change the PN. Regarding OS, the model estimated risk was slightly higher than the observed, with an O/E ratio of 0.73 (CI 95% 0.57-0.93) and 0.60 (0.49-0.74) at 2- and 5- years and calibration slope of 1.07 (0.83-1.31). For PFS, the O/E ratio was 0.48 (0.40-0.58) at 2 years and 0.50 (0.42-0.58) at 5 years, and a calibration slope of 0.89 (0.66-1.12). Both for OS and PFS rescaling of baseline hazard and model coefficients resulted in a better calibrated model. Model was validated also after missing data imputation, via MCE (m=5) showing consistent estimates. Conclusions: We validated the prognostic nomogram developed by Fakhry et al. in a large independent European cohort, demonstrating its applicability. These real-life data may represent the benchmark for the design of new prospective clinical trials. Pts characteristics Age (years) Median age 62.1 (37.7-100) ECOG PS 0 1 2 Missing 508 - 64.6% 239 - 30.4% 35 - 4.5% 4 - 0.5% Staging (AJCC 7th edition) I-II III IVa IVb Missing 15 - 1.9% 89 - 11.3% 576 - 73.3% 94 - 12.0% 12 – 1.5% p16 status Negative Positive Missing 249 – 31.7% 427 - 54.3% 110 - 14.0% Marital status Married Single/Divorced Missing 451- 57.4% 200 – 25.4% 135 – 17.2% Education High school (or less) University Missing 463 - 58.9% 169 – 21.5% 154 – 19.6% Smoking History Never < 10 p/y > 10 p/y Missing 201 - 25.7% 98 - 12.5% 484 - 61.6% 3 – 0.2% Weight loss < 5% > 5% Missing 520 - 66.2% 143 - 18.2% 123 - 15.6% Anemia No Yes Missing 351 - 44.7% 174 – 22.1% 261 - 33.2%