Surveillance as Determinant of Long-Term Survival in Non-Transplanted Hepatocellular Carcinoma Patients

Abstract

Purpose: We aimed at assessing the impact of surveillance on long-term survival in HCC patients.

Methods: From the ITA.LI.CA database, we selected 1028 cases with long (≥5 years, LS group) and 2721 controls with short-term survival (<5 years, SS group). The association between surveillance and LS was adjusted for confounders by multivariable logistic regression analysis. Survival of surveilled patients was presented both as observed and corrected for the lead-time bias, and the comparison of survival between surveillance and no surveillance groups was also performed after balancing the baseline characteristics with inverse probability weights (IPW).

Results: LS patients were more frequently diagnosed under surveillance (p < 0.0001), and had more favorable baseline characteristics. Surveillance was an independent predictor of LS (OR = 1.413, 95% CI 1.195-1.671; p < 0.0001). The observed and the lead-time corrected survival of surveilled patients were significantly longer compared to the survival of not surveilled patients (p < 0.0001 and p = 0.0008, respectively). In IPW adjusted populations, no survival differences were demonstrated between the two groups (p = 0.30).

Conclusions: Surveillance, increasing early-stage diagnosis and applicability of curative treatments, is a fundamental determinant of long-term survival in HCC patients. A wide implementation of surveillance programs should be pursued in order to improve HCC patients' prognosis.

Keywords: cancer stage; hepatocellular carcinoma; long-term survival; surveillance; treatment.

Figure 1

Study flow chart. Selection of patients finally included in the case (long-term survivors—LS) and control (short-term survivors—SS) groups.

Figure 2

Kaplan–Meier survival curves comparing surveillance and no surveillance groups. (A) Observed survival of patients diagnosed under surveillance or with a casual/symptomatic diagnosis. Patients diagnosed under surveillance demonstrated a significantly longer survival (p < 0.0001). (B) Observed survival of patients with casual/symptomatic diagnosis compared to corrected survival in surveilled patients. Surveillance significantly improves prognosis of patients even after correction for the lead-time bias (p = 0.0008). (C) Comparison of survival between surveilled and not surveilled patients after adjustment for adjustment for confounders with IPW. The two groups of patients showed similar survival (p = 0.30).