Cancer, also known as malignant tumor or malignant neoplasm, is the name given to a collection of related diseases. In all types of cancer, some of the body's cells begin to divide abnormally without stopping and have the potential to invade surrounding tissues. In this work, we focus on estimating the parameters of a model which tries to describe the growth of a cancer tumor based on the available measurements of the tumor volume and on comparing the effectiveness with respect to the accuracy of the estimation of the various methods we have tested. The Gompertz function is used as the model basis, and our analysis aims to compute the growth rate and the plateau size of the tumor. The methods used to estimate these parameters are based on least squares, maximum likelihood and the Extended Kalman Filter (EKF). In this work, we present five different methods. The results show that, when the process and measurement noise characteristics are known, maximizing the joint probability density function of the observations using numerical integration to compute the probability density functions yields most times the best results. The methods based on the EKF also yield satisfactory results.

Using nonstandard, problem specific techniques can improve the estimation accuracy and best exploit the available data.

No abstract

No abstract

Several mathematical models have been developed to explain the growth of tumors and used to fit experimental or clinical data. Their predictive poweri.e. their ability to forecast the future growth on the basis of present knowledge-however, has been rarely explored. Here, we investigate whether a Hidden Markov Model (HMM) based on the well-established Gompertz tumor growth function with additive Gaussian noise could effectively be used to predict the future growth of experimental tumors. The idea behind this work is that one might achieve more accurate predictions if estimates of the unknown parameters of the HMM are used instead of those obtained by fits of the deterministic Gompertz model to the data. We use the principle of Maximum Likelihood