A group of researchers in the U.K. have found that whole body MR imaging may reduce the time required for staging and planning cancer treatment, based on two prospective trials.
Conducted on nearly 500 patients across 16 hospitals, the Streamline C and Streamline L trials compared whole body MR to multiple imaging techniques currently recommended by guidelines to assess patients recently diagnosed with colorectal cancer and non-small cell lung cancer, respectively.
“The current guidelines still recommend multiple techniques because, to date, there has been limited supportive data for WB-MR from large prospective multicenter studies such as the Streamline trials,” lead author Stuart Taylor, professor of medical imaging at University College London, told HCB News. “Until now, studies supporting the use of WB-MR have mainly been quite small, performed at one or two hospital sites only and used a small number of experienced radiologists to interpret the scans. Health care policy makers cannot use this level of evidence to recommend WB-MR. We hope the Streamline trials will change this.”
Determining the right course of treatment for cancer requires clinicians to know the size of a tumor and the extent to which is has spread to nearby lymph nodes, as well as other parts of the body. Standard pathways outlined by the National Health Service involve different imaging modalities, such as CT, PET/CT or focused MR, which vary in accuracy depending on the organ. As such, several appointments and follow-up exams can be required.
WB-MR, which can image the entire body in one hour or less, was found to reduce the average time for determining tumor size, and the extent of spreading, by five days for colorectal cancer patients and six for lung cancer patients.
While sensitivity and specificity of diagnosis for WB-MR did not vary from that of standard tests for both forms of cancer, it did reduce the time required for complete diagnostic exams from an average of 13 to eight days for colorectal cancer patients, and from 19 to 23 days for those with lung cancer. Costs also decreased from an average of £285 to £216 in the colorectal cancer trial and an average of £620 to £317 in the lung cancer trial.
MR technology, however, is lacking in availability compared to other imaging modalities, due to its high demand, according to Taylor. Many hospitals in the trials, for instance, were unable to make time on their MR scanners, forcing patients to receive exams at nearby practices. Additionally, eight of the 16 hospitals in the colorectal trial lacked the infrastructure necessary to perform WB-MR, as did 11 of 16 in the lung cancer trial.
Taylor and his colleagues plan to continue their studies on WB-MR, having applied for a grant to research its use in staging breast cancer. They also plan to work with Imperial College London in studying the influence of AI and machine learning applications on interpretations of such scans.
“WB-MR consists of many images (as several sequences are acquired through the whole body),” he said. “Thus, they are quite difficult and time-consuming to interpret. AI and machine learning may be able to highlight areas of abnormality on the scans for radiologists to check, speeding up the time it take to interpret the study and potentially increasing accuracy by reducing perceptual errors.”
To perform their research, a multidisciplinary panel compared the diagnostic accuracy and efficiency of whole body MR to that of standard imaging tests recommended by the National Institute for Health and Care Excellence (NICE). The findings of the latter were used to make a first-treatment decision, and then images from the WB-MR were evaluated. If the NICE tests indicated a need for further testing, it was carried out, as it allowed the panel to determine if the first treatment decision would have been different, based on WB-MR results.
The final decision took into consideration the findings of WB-MR and standard testing, with patients followed up after 12 months to assess the accuracy of both, and retrospectively evaluate what the optimal treatment decision should have been. The final multidisciplinary panel treatment decision in the colorectal cancer trial was 95 percent for standard investigations and WB-MR. For the lung cancer trial, it was 99 percent and 98 percent, respectively.
The researchers caution that the findings may not be relevant to tumors in other areas of the body, as they are only specific to colorectal and non-small cell lung cancer, and that sensitivity in detecting the spread of cancers, including the development of secondary tumors and the spread of lymph nodes, is low when using both standard imaging methods and WB-MR. Another limitation is the fact that the waiting times in the trials may not represent those of other U.K. providers or hospitals in other countries.
The findings were published in The Lancet Gastroenterology & Hepatology
and The Lancet Respiratory Medicine