Discriminatory CT-textural features in splenic infiltration of lymphoma versus splenomegaly in liver cirrhosis versus normal spleens in controls and evaluation of their role for longitudinal lymphoma monitoring

PURPOSE

To find CT-texture analysis (CTTA) features for the discrimination of splenomegaly due to diffuse lymphoma involvement and liver cirrhosis versus normal-sized spleens in controls and to assess their potential role for longitudinal lymphoma monitoring.

MATERIAL AND METHODS

We had retrospectively identified 74 subjects with diffuse splenic involvement due to lymphoma (n = 29) and liver cirrhosis (n = 30), and healthy controls (n = 15), who underwent contrast-enhanced abdominal CT between August 2013 and October 2017. CTTA evaluation included heterogeneity, intensity, average, deviation, skewness, entropy of co-occurrence, number non-uniformity (NGLDM) and entropy NGLDM. A greater than 50% reduction of spleen volume after chemotherapy was considered proof for splenic involvement.

RESULTS

There were significant differences of splenic CTTA-values before and after treatment of patients with lymphoma, including mean of entropy(p < .001), uniformity of average(p < .001), uniformity of deviation(p = .002) and entropy of skewness(p < .001). Significant differences of splenic CTTA-values in subjects with lymphoma vs. healthy controls were found for mean intensity(p < .001), mean average(p < .001), and entropy of deviation(p < .001). No significant differences in splenic CTTA-values were found in subjects with lymphoma that reached complete remission vs. controls. Splenic CTTA values mean intensity(p = .002) and mean average(p = .004) were significantly different between subjects with untreated lymphoma and subjects with liver cirrhosis. At end-of-treatment all lymphomas reached complete remission. Entropy/uniformity of heterogeneity(p < .001), mean intensity(p = .007), mean average (p = .007), uniformity of average(p = .008) and mean/entropy/uniformity of skewness(p = .001) measured at this time differed significantly from baseline.

CONCLUSIONS

CTTA features in subjects with splenomegaly due to lymphoma and liver cirrhosis differ significantly from those of healthy controls and can be also used for monitoring lymphoma treatment. Quantitative CTTA features may increase the accuracy of diagnosing causes of splenomegaly.