Towards A Better Paradigm for Head and Neck Cancer Treatment Applying Artificial Intelligence. HNC-TACTIC.

Towards A Better Paradigm for Head and Neck Cancer Treatment Applying Artificial Intelligence: an International Cohort Study of Electronic Health Records. HNC-TACTIC.

This will be an international, multicenter, retrospective, observational, and data-driven study using secondary data captured in EHRs. The extraction of the data captured in the EHRs will be performed with SAVANA's EHRead®, an innovative data-driven system based on Natural Language Processing (NLP) and machine learning. For all patients, the Index Date is defined as the timepoint within the study period when they fulfill ALL inclusion criteria and no exclusion criteria. Follow-up comprises the period between Index Date and the last EHR available within the study period. Additional variable-specific time windows may be considered to optimize data collection.

The present study aims to describe the clinical characteristics of patients with HNSCC in a real-world setting by analyzing readily available information in the Electronic Health Records (EHRs). This study will gain a deep insight of the clinical characteristics and real-world outcomes of patients with all stages (early, locally advanced, and metastatic) of HNSCC. It will focus on developing two predictive models to apply in the clinical setting, one for electing patients with high-risk of recurrence after radical treatment, and the second one for selecting recurrent or metastatic patients who could benefit from immunotherapy.

To achieve the proposed study objectives we will use SAVANA´s EHRead® (11-15), a technology that applies Natural Language Processing (NLP) (16) and machine learning to extract, organize, and analyze the unstructured clinical information jotted down by health professionals in patients' EHRs.

Primary objectives

• To develop a predictive model based on dynamic risk stratification (DRS) for the risk of recurrence or disease progression following a primary curative treatment in HNSCC patients with early and locally advanced disease.
• To develop a predictive model based on dynamic risk stratification (DRS) aimed at identifying patients' features that predict long-term survival after immunotherapy in recurrent and metastatic HNSCC patients. Secondary objectives
• To describe median OS by primary tumor location (oral cavity, oropharynx, larynx, and hypopharynx) in HNSCC patients after stratification for prognostic factors, including tumor stage and treatment.
• To describe the demographics, clinical characteristics, and treatment of patients with HNSCC in early and locally advanced stages of the disease.
• To describe the patterns of follow-up in patients with HNSCC in early and locally advanced stages of the disease.
• Departments in charge
• Number of visits
• Imaging and anatomopathological tests
• Recurrence detected by physical examination.
• To evaluate the impact of treatments on patients with locally advanced stages of the disease.
• Patients' early and late toxicity to the treatment, comparing between radiotherapy (+/-cisplatin or cetuximab) vs surgery and post-operative r< radiotherapy (+/- cisplatin).
• Healthcare resource utilization (HCRU), including medical visits, diagnostics, and hospitalizations.
• To compare OS in locally advanced HNSCC patients (including both HPV+ and HPV- oropharyngeal patients) treated with cisplatin-radiotherapy vs cetuximab-radiotherapy and treated with surgery vs. conservative treatment.
• To compare the demographic and clinical characteristics of exceptional responders and poor responders (based on recurrence and long-term survival). This analysis will be performed independently for HPV+ and HPV- oropharyngeal patients.

Inclusion Criteria:

• Patients ≥18 years old.
• Patients diagnosed with HNSCC
• For selected exploratory analyses, patients diagnosed with nasopharynx, paranasal sinus, and salivary gland tumors.

Exclusion Criteria:

• Patients with follow-up of less than 6 months, except if deceased (any cause) in the 6 months after HNSCC diagnosis