Oleg Kshivets , MD, PhD
Surgery Department, Kaluga Cancer Clinical Center, Russia
PRECISE PREDICTION OF 5-YEAR
SURVIVAL OF LUNG CANCER PATIENTS
AFTER RADICAL SURGERY
ABSTRACT
Precise Prediction of 5-Year Survival of Lung Cancer Patients after Radical Surgery
Oleg Kshivets
OBJECTIVE: This study aimed to determine homeostasis and tumor factors for 5-year survival (5YS) of non-
small cell lung cancer (LC) patients (LCP) (T1-4N0-2M0) after complete en block (R0)
lobectomies/pneumonectomies (LP).
METHODS: We analyzed data of 665 consecutive LCP (age=57.5±8.3 years; tumor size=4.4±2.4 cm) radically
operated and monitored in 1985-2015 (m=575, f=90; lobectomies=423, pneumonectomies=242, combined LP with
resection of trachea, carina, atrium, aorta, VCS, vena azygos, pericardium, liver, diaphragm, ribs, esophagus=180;
only surgery-S=524, adjuvant chemoimmunoradiotherapy-AT=141: CAV/gemzar + cisplatin + thymalin/taktivin +
radiotherapy 45-50Gy; T1=237, T2=248, T3=125, T4=55; N0=419, N1=130, N2=116, M0=665; G1=163, G2=199,
G3=303; squamous=377, adenocarcinoma=243, large cell=45; early LC=132, invasive LC=533. Multivariate Cox
modeling, clustering, SEPATH, Monte Carlo, bootstrap and neural networks computing were used to determine any
significant dependence.
RESULTS: Overall life span (LS) was 2114.8±1685 days and cumulative 5YS reached 69.6%, 10 years – 61.2%,
20 years – 43.1%. 416 LCP lived more than 5 years without cancer (LS=3041.4±1472.5 days). 193 LCP died because
of LC (LS=559.6±373.5 days). AT significantly improved 5YS (65.1% vs. 34.3%) (P=0.00001 by log-rank test) only
for LCP with N1-2. Cox modeling displayed (Chi2=290.78, df=13, P=0.000) that 5YS of LCP significantly depended
on: phase transition (PT)“early-invasive LC”, PT N0-N12, histology, G, blood cell subpopulations, cell ratio factors
(ratio between blood cells subpopulations and cancer cells-CC), prothrombin index, heparin tolerance, recalcification
time, glucose, AT (P=0.000-0.035). Neural networks, genetic algorithm selection and bootstrap simulation revealed
relationships between 5YS and PT N0-N12 (rank=1), PT “early-invasive LC” (rank=2), lymphocytes (3), segmented
neutrophils (4), tumor size (5), AT (6), T1-4 (7), ESS (8), prothrombin index (9), glucose (10), thrombocytes/CC (11),
healthy cells/CC (12), lymphocytes/CC (13), erythrocytes/CC (14). Correct prediction of 5YS was 100% by neural
networks computing (error=0.000; area under ROC curve=1.0).
CONCLUSIONS: 5YS of LCP after radical procedures significantly depended on: tumor characteristics, blood cell
circuit, cell ratio factors, hemostasis system and AT.
Results of Kohonen Self-Organizing Neural Networks
Computing in Prediction of Lung Cancer Patients Survival
after Complete Lobectomies/Pneumonectomies (n=609):