Metastases Growth Patterns in vivo—A Unique Test Case of a Metastatic Colorectal Cancer Patient

Abstract

Colorectal cancer (CRC) is one of the most common causes of cancer-related mortality worldwide. Most cases of deaths result from metastases, assumed to be shed, in many cases, before disease detection. Providing reliable predictions of the metastases' growth pattern may help planning treatment. Available mathematical tumor growth models rely mainly on primary tumor data, and rarely relate to metastases growth. The aim of this work was to explore CRC lung metastases growth patterns. We used data of a metastatic CRC patient, for whom 10 lung metastases were measured while untreated by seven serial computed tomography (CT) scans, during almost 3 years. Three mathematical growth models-Exponential, logistic, and Gompertzian-were fitted to the actual measurements. Goodness of fit of each of the models to actual growth was estimated using different scores. Factors affecting growth pattern were explored: size, location, and primary tumor resection. Exponential growth model demonstrated good fit to data of all metastases. Logistic and Gompertzian growth models, in most cases, were overfitted and hence unreliable. Metastases inception time, calculated by backwards extrapolation of the fitted growth models, was 8-19 years before primary tumor diagnosis date. Three out of ten metastases demonstrated enhanced growth rate shortly after primary tumor resection. Our unique data provide evidence that exponential growth of CRC lung metastases is a legitimate approximation, and encourage focusing research on short-term effects of surgery on metastases growth rate. SIGNIFICANCEProviding reliable predictions of the metastases' growth pattern using mathematical models may help determining the optimal treatment plan that fits a given patient best and maximizes the probability of cure.

Metastasis Initiation Precedes Detection of Primary Cancer—Analysis of Metastasis Growth in vivo in a Colorectal Cancer Test Case

“…These metastases were first detected 2.6 years after primary tumor resection, and 1.7 years after a liver metastasis was resected in a second operation. This case is different from the former (Hochman et al, 2019 ) in the location of the primary tumor—sigmoid colon in this case, and rectal in the former case, and also in the fact that here there were two metastatic locations (liver and lungs), and two operations were conducted. These distinctions imply a different type of lung metastases, with a possibly different route of metastatic spread, which may induce a different course of the natural history of metastases.…”

“…Based on the measurement data obtained, we wanted to fit a growth model (exponential, logistic, or Gompertzian) for each of the metastases and assess growth rate parameter values using the same methods described in Hochman et al ( 2019 ).…”

“…Lately, we have published an analysis of data from a rare test case of a metastatic CRC patient, with untreated growth of 10 lung metastases repeatedly measured over 3 years (Hochman et al, 2019 ). We have shown that exponential growth can be approximated to all metastases and that metastases were initiated at least 8–11 years before the primary disease detection.…”

“…This is compared to the former case, where at the time of disease detection there was a 6-cm 3 tumor in the rectum, a colonic polyp, and at least eight lung metastases were already detectable. Here, we re-examine the prediction of early metastasis onset time, not only for metastases observed at first detection of the primary (as in Hochman et al, 2019 ) but also for metastases that were occult at the time of disease detection and observed only 2 years afterward. In this case of late-detected metastases, we also faced a more difficult task as we wanted to study the effects of the two surgeries on metastasis growth.…”

Changing Oncology Treatment Paradigms in the COVID-19 Pandemic

“…13 At the convergence of mathematical oncology, cancer cell biology, and immunology, there may be a meeting point that will help solve modern and urgent problems in practical cancer treatment by informing better timing of cancer treatment. [14][15][16] An adaptive therapy strategy and clinical trial design, based on prediction models of the evolution of cancer cell growth and response to drugs, are already being used for clinical trials in metastatic castrate-resistant prostate cancer. 17 There is opportunity to leverage the evolution of the field of medical oncology as well to better incorporate perspectives from experts in mathematical oncology and bioengineering to likewise address this important aspect colon cancer biology into the way we design trials with therapeutic intent, with arms that reflect appropriate changes based on changes in tumors identified using serial liquid biopsies (e.g., circulating tumor DNA, pharmacogenomically profiled circulating tumor cells) and other biologic tools.…”