Abstract 2091: DeepBTS: Prediction of recurrence-free survival of non-small cell lung cancer using time-binned deep neural network

Author(s):  
Ho Jung An ◽  
Bora Lee ◽  
Sang Hoon Chun ◽  
Ji Hyung Hong ◽  
In Sook Woo ◽  
...  
2020 ◽  
Vol 50 (11) ◽  
pp. 1306-1312
Author(s):  
Atsushi Kagimoto ◽  
Yasuhiro Tsutani ◽  
Yoshinori Handa ◽  
Takahiro Mimae ◽  
Yoshihiro Miyata ◽  
...  

Abstract Objectives This study aimed to determine the characteristics, ground glass opacity ratio and prognosis of patients with clinical N0 non-small cell lung cancer tumours exceeding 30 mm in size. Methods Patients with clinical N0 non-small cell lung cancer and total tumour size >30 mm on preoperative computed tomography who underwent complete resection with lobectomy between January 2007 and December 2017 were included. The patients were divided into three groups: pure solid tumour, low ground glass opacity ratio (1–39%) tumour and high ground glass opacity ratio (≥40%) tumour. The cut-off line was determined based on the recurrence rate for every 10% ground glass opacity ratio. Results Among the 227 study patients, 129 (56.8%) had a pure solid tumour, 54 (23.8%) had a low ground glass opacity ratio tumour and 44 (19.4%) had a high ground glass opacity ratio tumour. Three-year recurrence-free survival was significantly shorter in patients with a pure solid tumour (57.4%) than in patients with a low ground glass opacity ratio (74.5%; P = 0.009) or a high ground glass opacity ratio tumour (92.1%; P < 0.001). Multivariable analysis showed that ground glass opacity ratio was a significant independent prognostic factor for recurrence-free survival (hazard ratio, 0.175; P = 0.037). Conclusion Pure solid tumours comprised a large proportion of non-small cell lung cancer tumours >30 mm in size and their prognosis was poor. The presence of ground glass opacity and their relative proportion affect prognosis in patients with clinical N0 non-small cell lung cancer tumours >30 mm in size, similar to those with small-sized tumours.


Sign in / Sign up

Export Citation Format

Share Document