Title

Delving into Survival of Colorectal Cancer - Long-term Survival, Younger Populations, and Prediction Models

Date of Award

Spring 2021

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Public Health

First Advisor

Fuchs, Charles

Abstract

Colorectal cancer (CRC) is one of most common cancers in the United States and is a longstanding health threat to men and women. Although survival rates after CRC diagnosis have markedly increased since 1975, likely due to advances in early screening and treatment, there are three important aspects demanding further investigation: 1) the long-term survival and causes of death after CRC diagnosis, 2) the differences of survival of early-onset CRC (defined as age <50 years) compared to CRC diagnosed at older ages, and 3) the development of novel comprehensive prediction models for personalized survival estimates. In this dissertation, I delve into the above research areas that may advance our knowledge in CRC survival.In Chapter 1, I investigated long-term survival rates and causes of death among men and women diagnosed with colorectal cancer in three cohorts of U.S. health professionals - Nurses’ Health Study, Nurses’ Health Study II, and Health Professionals Follow-up Study. After over three decades of follow-up, among male CRC survivors, the 30-year cumulative overall mortality rate was 93.2%, and cancer-specific mortality rate was 37.0%. Due to longer follow-up of women, we observed 86.5% for 35-year cumulative overall mortality rate, and 39.6% for cancer-specific mortality rate. Regarding causes of death, given that cancer-specific mortality rate increased by <4% between 10 to 30- or 35-year following diagnosis (among men, from 35.2% to 37.0%; among women, from 35.8% to 39.6%), CRC patients surviving more than 10 years after diagnosis are unlikely to die from that malignancy and other diseases should receive more attention in treatment of survivors beyond 10 years. This is good news for CRC patients. In Chapter 2, given that the incidence rate of early-onset CRC rapidly increased by 19.7% from 2012 to 2017, it is of great importance to look into survival after diagnosis of early-onset CRC. Using the National Cancer Database, I assessed the survival differences of early-onset CRC and later-onset CRC (defined as ages 51-55 years in this study), and explore the heterogeneity of survival within early-onset CRC. Early-onset CRC patients experienced a modestly inferior overall survival for all years in unadjusted analysis, compared to later-onset CRC. However, following adjustment for other predictors for mortality, most notably disease stage, early-onset CRC patients had a lower risk of death compared to subjects diagnosed between ages 51-55 (adjusted hazard ratio: 0.95 [95% CI, 0.93-0.96]). The survival advantage appeared greatest for patients diagnosed at ages 35-39 and stages I-II, and was absent among those diagnosed at 25 years or younger and stages III-IV. The survival advantage of early-onset CRC for younger patients should be interpreted cautiously, given that the advantage has small magnitude and is heterogeneous by age and stage. Further study is needed to understand the underlying heterogeneity of survival within early-onset patients by age and stage. In Chapter 3, I specially focused on developing novel prediction models for colon cancer. Although CRC is a term that combines colon cancer and rectal cancer, colon cancer accounts for >70% of CRC. Using a cohort derived from a large multiple-center randomized trial CALGB 89803, I developed visual nomograms of prediction models by clinical, pathologic, diet, and other lifestyle characteristics of colon cancer patients. These models can help clinicians and their patients quickly estimate 5-year disease-free survival and overall survival. Additionally, we observed that patients with healthy diet, weight control, and higher levels of physical activity are predicted to have superior survival rates and decreased risk of cancer recurrence and mortality compared with most patients. Thus, prediction models incorporating diet and lifestyle factors may help clinicians and patients evaluate personalized survival expectations and inform diet and lifestyle modification as part of cancer treatment.

This document is currently not available here.

Share

COinS