ESTRO 2022 - Abstract Book

S1061

Abstract book

ESTRO 2022

The combination of systemic and local treatments without frontline immunotherapy setting showed encouraging results, with a 5-years OS of 31%.

PO-1256 Multi-domain automated lung segmentation for inflammatory lung disease (ILD) detection

A. Hope 1 , C. McIntosh 2 , M. Welch 3 , S. Kandel 4 , T. Purdie 5 , T. Tadic 2 , T. Patel 2

1 Princess Margaret Cancer Center / University of Toronto, Radiation Medicine Program / Radiation Oncology, Toronto, Canada; 2 Princess Margaret Cancer Center, Radiation Medicine Program, Toronto, Canada; 3 Princess Margaret Cancer Center, Data Science, Toronto, Canada; 4 University Health Network, Joint Department of Medical Imaging, Toronto, Canada; 5 Princess Margaret Cancer Center / University of Toronto, Radiation Medicine Program / Medical Biophysics, Toronto, Canada Purpose or Objective ILD can predispose patients to high risk of pulmonary complications or even death following high dose radiation therapy. Unfortunately, not all patients with ILD are known at the time of radiation treatment decision. Automated methods to detect ILD on diagnostic and/or planning CTs would provide multiple checks to ensure patient safety, but requires high quality lung segmentation of both patients with and without ILD as a prerequisite. Materials and Methods Using a training set (TRN) of 214 radiation planning computed tomographic (CT) images from NSCLC patients including cases with and without ILD, a convolutional neural net (CNN) was trained to automatically segment the lungs within these scans. We trained a CNN based on the U-Net topology but using 3D convolution filters in place of the traditional 2D. Due to memory limitations all images resampled to 256x256x128. After training, two validation datasets were generated composed of radiation treatment planning CTs from NSCLC patients (VAL1, n=24) and a set of diagnostic thoracic CT images (VAL2, n=100[CM1] ). All patients in VAL2 were further labeled by the same radiologist as to whether ILD was radiographically present or absent. Test characteristics of the CNN were calculated using the Dice metric on VAL1, and qualitative inspection on VAL2. Results After training, the CNN demonstrated Dice of 0.96 on VAL1 and strong qualitative agreement on VAL2, uniquely demonstrating that a CNN can be trained on RT planning CTs to segment both planning and diagnostic imaging. The CNN takes on average 4.5 seconds to segment a novel image with roughly half that time dedicated to reading the image from disk. ILD was present in 20% of cases in VAL2. Conclusion A CNN has been developed that can rapidly segment radiation treatment planning CT or diagnostic CTs to enable downstream automation of a system to identify patients at high risk of having pre-existing ILD. After prospective validation, this tool and similar tools could be incorporated into radiation treatment planning systems to automatically alert clinicians about high-risk patients that might have proceeded to treatment planning without having pre-existing ILD identified. G. Corrao 1 , M. Franchi 2 , G. Marvaso 1 , M. Zaffaroni 3 , M. Pepa 4 , S. Volpe 1 , M.G. Vincini 3 , G. Piperno 3 , A. Ferrari 5 , B.A. Jereczek-Fossa 1 1 IEO, European Institute of Oncology IRCCS; University of Milan, Division of Radiation Oncology; Department of Oncology and Hematoncology, Milan, Italy; 2 University of Milan-Bicocca, Department of Statistics and Quantitative Methods, Milan, Italy; 3 IEO, European Institute of Oncology IRCCS, Division of Radiation Oncology, Milan, Italy; 4 IEO, European Institute of Oncology IRCCS, Division of Radiation Oncology , Milan, Italy; 5 IEO, European Institute of Oncology IRCCS,, Division of Radiation Oncology, Milan, Italy Purpose or Objective Healthcare administrative datasets represent a valuable source for real-life data analysis. In the current work data on more than 10 million individuals resident in Lombardy Region (Northern Italy) were considered. Primary aim is to compare safety and effectiveness in a non-small cell lung cancer (NSCLC) patients who received different patterns of first-line systemic therapy with or without RT. Materials and Methods Diagnostic ICD-9-CM codes were used for identifying all patients with a new diagnosis of lung cancer between 2012 and 2019. Among these, patients who started a systemic non-chemotherapy as first-line of treatment for advanced NSCLC alone or in combination with radiotherapy (RT). Since the same code is applied to NSCLC and SCLC, systemic treatment (ST) considered were tyrosine-kinase inhibitors (TKI) or pembrolizumab as drugs exclusively administered in NSCLC. RT treatments were limited to SBRT and IMRT, in order to select patients with a better expected prognosis. Patients were followed from the date of first-line treatment start until 31 st December 2020. Overall survival (OS) was estimated by using the Kaplan-Meier estimator and differences between groups were compared using the log-rank test. Hazard ratios, along with 95% confidence intervals (CI) were estimated using Cox proportional hazards models adjusted for sex, age, year of first-line treatment start and a cancer multimorbidity score. Analyses were stratified by type of first-line ST and diagnosis of brain metastasis. PO-1257 Impact of modern RT in advanced NSCLC: an exploratory real-life investigation from Lombardy