CT Image-based Decision Support System for Categorization of Liver Metastases Into Primary Cancer Sites

Rationale and Objectives

This study aimed to provide decision support for the human expert, to categorize liver metastases into their primary cancer sites. Currently, once a liver metastasis is detected, the process of finding the primary site is challenging, time-consuming, and requires multiple examinations. The proposed system can support the human expert in localizing the search for the cancer source by prioritizing the examinations to probable cancer sites.

Materials and Methods

The suggested method is a learning-based approach, using computed tomography (CT) data as the input source. Each metastasis is circumscribed by a radiologist in portal phase and in non-contrast CT images. Visual features are computed from these images, combined into feature vectors, and classified using support vector machine classification. A variety of different features were explored and tested. A leave-one-out cross-validation technique was conducted for classification evaluation. The methods were developed on a set of 50 lesion cases taken from 29 patients.

Results

Experiments were conducted on a separate set of 142 lesion cases taken from 71 patients with four different primary sites. Multiclass categorization results (four classes) achieved low accuracy results. However, the proposed system was found to provide promising results of 83% and 99% for top-2 and top-3 classification tasks, respectively. Moreover, when compared to the experts' ability to distinguish the different metastases, the system shows improved results.

Conclusions

Automated systems, such as the one proposed, show promising new results and demonstrate new capabilities that, in the future, will be able to provide decision and treatment support for radiologists and oncologists, toward more efficient detection and treatment of cancer.

Introduction

Computed tomography (CT) is one of the most common modalities used for the detection, diagnosis, and follow-up of liver lesions (1), with the images acquired before and after intravenous injection of a contrast agent (Iohexol 350 mg/mL up to 2 mL/kg) and oral contrast agent (Iohexol 350 mg/mL 52 mL diluted in 2 L water). Radiologists usually detect and diagnose liver lesions based on the different density of the lesions at different time points in the scan.

The liver is one of the most common organs to develop metastases. Metastatic liver lesions can be derived from different primary cancer types, including melanoma, breast cancer, colon cancer, neuroendocrine, pancreatic cancer, and others. Not infrequently when liver metastases are detected, the radiologists do not know the location of the primary site of cancer. In these cases, a broad search for the source of the primary tumor is performed. In some cases, the source of the cancer remains unknown, despite a thorough physical examination including breast, pelvic and rectal examination, basic laboratory tests, urinalysis, X-rays, CT scans, and other imaging studies. Knowing the source of the primary tumor is important to help localize the search, for example, colonoscopy for colon carcinoma, gastroscopy for gastric carcinoma, etc, and to treat it accordingly.

Finding the primary site can be a time-consuming task requiring multiple examinations. The difficulty of this task highlights the need for assistive computerized tools in this domain. Several studies that categorize metastatic tumours by taking biopsies and using molecular gene expression profiling to predict the tissue of origin can be found 2, 3, 4. Segal et al. (5) showed that dynamic imaging traits in CT examinations systematically correlate with the global gene expression programs of primary human liver cancer. To the best of our knowledge, no methods to identify the source of the metastasis based on its appearance, directly from the CT image scan, currently exist.

The current work focuses on such a solution. We present an automated algorithm for categorization of liver metastases to their primary cancer sites using CT examinations following existing radiological workflow procedures. The method proposed is presented in the Methods section. Experimental results are shown in the Experiments and Results section. The experiments include the derivation of the final classification framework used in this work. In the Discussion section, a discussion of the results is presented followed by a conclusion of the work.