Prediction of response to neoadjuvant immunotherapy in patients with esophageal squamous cell carcinoma by a rapid breath test

Zheng Liu 1,2, Qi Huang 3, Zhiwei Rong 4, Yipei Yu 4, Peiyu Wang 1,2, William C Cho 5, Teng Mu 3, Jilun Li 3, Jia Zhao 3, Shaodong Wang 1,2, Xiangnan Li 3, Yan Hou 4, Mantang Qiu 1,2

1. Department of Thoracic Surgery, Peking University People’s Hospital, Beijing 100044, China
2. Thoracic Oncology Institute, Peking University People’s Hospital, Beijing 100044, China
3. Department of Thoracic Surgery, the First Affiliated Hospital of Zhengzhou University, Zhengzhou 450003, China
4. Department of Biostatistics, School of Public Health, Peking University, Beijing 100191, China
5. Department of Clinical Oncology, Queen Elizabeth Hospital, Hong Kong SAR, China 

Poster PDF

View presentation

Abstract: 

Neoadjuvant immunotherapy combined with chemotherapy have made great progress in the treatment of Esophageal squamous cell carcinoma (ESCC). There is no satisfying biomarker has been found to predict the response to neoadjuvant immunotherapy. Here, we present a novel technique to predict ESCC patient’s response to neoadjuvant immunotherapy by testing volatile organic compounds (VOCs) in exhaled breath. We screened 82 ESCC patients’ clinical data and VOC data. Becker criteria for tumor regression grading was used to assess treatment response of patients who received surgery. And patients who never underwent surgery were evaluated according to the targeted lymph node response rate based on iRECIST. We established the platform called high-pressure photon ionization time-of-flight mass spectrometry (HPPI-TOFMS) to test base line VOCs. It only took 60s to analyze each sample by HPPI-TOFMS. A total of 82 patients were enrolled, including 58 responders achieved complete response or partial response and 24 non-responders with stable disease or progressive disease. All enrolled ESCC patients were treated with platinum-paclitaxel/docetaxel chemotherapy plus one kind of anti-PD-1 immune checkpoint inhibitor. After smoothing, baseline correction, peak detection, peak alignment, and peak area calculation, 264 features were extracted from each sample. With these features, breath test successfully predicted neoadjuvant immunotherapy response with area under curve (AUC) of 0.73±0.09. By integrating breath test with routine blood test data, the prediction performance improved. In the validation set, the integrated model reached a sensitivity of 83.34%±23.89%, specificity of 80.0%±9.61%, accuracy of 80.0%±8.00%, and AUC of 0.82±0.10. Given the advantages of high acceptability, tolerability, noninvasiveness, low cost, and high accuracy, breath test based on HPPI-TOFMS is a promising triage test for ESCC patients and a useful tool for clinical decision making.