As a metastasis finder, whole-body MRI wins some, loses some

Whole-body diffusion-weighted MRI (DWI) very nearly matches conventional imaging when it comes to finding metastases in the liver and bones of cancer patients. That’s the good news for whole-body DWI.

The bad news: On its own, it’s lousy at finding metastases in the lymph nodes and respiratory system.

So found Samir Mustaffa Paruthikunnan, MD, and colleagues at Manipal University in Karnataka, India, when they compared whole-body DWI with sectional CT, MRI and nuclear scintigraphy.

The American Journal of Roentgenology published their study online July 5.

The team performed all of the above imaging exams on 51 adult patients (26 women, 25 men) with known malignancies, then followed up with the patients for one year.

They had two radiologists separately assess the DWI images while a senior-level rad assessed the conventional images, classifying the metastatic lesions into one of four anatomic regions—liver, lung, skeletal system and lymph nodes.

The researchers based their reference standard on histopathologic confirmation or clinical follow-up of the metastases.

Whole-body DWI came in only slightly behind the conventional modalities at detecting hepatic and skeletal metastases, hitting 89.7 percent accuracy with the former and 85.2 percent for the latter, and the differences in score were not statistically significant.

However, the DWI technique was significantly inferior to the regional modalities at finding pulmonary and lymph-node metastases.

It was particularly weak with lung growths, where its sensitivity was 33.3 percent and its accuracy 60.8 percent. Here the conventional modalities achieved 100 percent sensitivity, specificity and accuracy, all three.

The authors note as their main limitation their hospital’s lack of FDG PET-CT, suggesting future comparison studies include that modality as well.

They conclude that whole-body DWI “can be used for screening hepatic and skeletal metastases, but its reliability as the sole imaging sequence for the detection of lymph nodal and pulmonary metastases is poor and, at present, it cannot replace conventional imaging modalities.”