Clinical correlates of CT imaging-derived phenotypes among lean and overweight patients with hepatic steatosis

The objective of this study is to define CT imaging derived phenotypes for patients with hepatic steatosis, a common metabolic liver condition, and determine its association with patient data from a medical biobank. There is a need to further characterize hepatic steatosis in lean patients, as its epidemiology may differ from that in overweight patients. A deep learning method determined the spleen-hepatic attenuation difference (SHAD) in Hounsfield Units (HU) on abdominal CT scans as a quantitative measure of hepatic steatosis. The patient cohort was stratified by BMI with a threshold of 25 kg/m2 and hepatic steatosis with threshold SHAD ≥  − 1 HU or liver mean attenuation ≤ 40 HU. Patient characteristics, diagnoses, and laboratory results representing metabolism and liver function were investigated. A phenome-wide association study (PheWAS) was performed for the statistical interaction between SHAD and the binary characteristic LEAN. The cohort contained 8914 patients—lean patients with (N = 278, 3.1%) and without (N = 1867, 20.9%) steatosis, and overweight patients with (N = 1863, 20.9%) and without (N = 4906, 55.0%) steatosis. Among all lean patients, those with steatosis had increased rates of cardiovascular disease (41.7 vs 27.8%), hypertension (86.7 vs 49.8%), and type 2 diabetes mellitus (29.1 vs 15.7%) (all p < 0.0001). Ten phenotypes were significant in the PheWAS, including chronic kidney disease, renal failure, and cardiovascular disease. Hepatic steatosis was found to be associated with cardiovascular, kidney, and metabolic conditions, separate from overweight BMI.

Supplementary Table S1: Uncorrected p-values from pairwise chi-square testing for clinical and demographic categorical characteristics for mild to severe steatosis.Age, BMI, LMA, and SHAD values listed for each group are in the format of median [interquartile range].A higher value for SHAD corresponds to increased hepatic steatosis.P-values were determined using the Kruskal-Wallis test for continuous fields (age) and chisquare for categorical fields (sex, race, CVD, HTN, T2DM, HLD) and were corrected using the Benjamini-Hochberg method for a significance threshold of p<0.05.P-values were not calculated for BMI, LMA, and SHAD because the groups were dichotomized by these quantities.Lean patients have 18.5 kg/m 2 ≤BMI<25 kg/m 2 .Patients with steatosis have SHAD≥10 HU or LMA<40 HU.BMI=body mass index, LMA=liver mean attenuation, SHAD=spleen-hepatic attenuation difference, CVD=cardiovascular disease, HTN=hypertension, T2DM=type 2 diabetes mellitus, HLD=hyperlipidemia.
Supplementary Table S3: Chi-squared and p-values from pairwise chi-square testing for clinical and demographic categorical characteristics, considering only moderate-to-severe steatosis. Lean Coefficients and p-values are from the PDFF x LEAN statistical interaction term of the logistic regression model controlling for age, sex, and race.LEAN is a binary characteristic indicating that a patient has 18.5 kg/m 2 ≤BMI<25 kg/m 2 .BMI=body mass index, PDFF=proton density fat fraction.Supplementary FigureS6: Phenome-Wide Association Study of the statistical interaction between the continuous characteristics SHAD and BMI (SHAD x BMI).Each coded phenotype is shown as a triangle and further grouped by color as indicated in the legend at the bottom.Upward pointing triangles indicate a positive association with SHAD x BMI, while downward pointing triangles indicate a negative association.The blue dashed horizontal line on the graph indicates the level of statistical significance with Bonferroni multiple-comparison correction (p<7.5x10-5 ).Phenotypes above the line are those that show a statistically significant association with SHAD x BMI.The seven statistically significant phenotypes in order of statistical significance are the following: gout, gout and other crystal arthropathies, renal failure, chronic renal failure (CKD), gouty arthropathy, thrombocytopenia, acute renal failure.BMI=body mass index, SHAD=spleen-hepatic attenuation difference.

Table S2 :
Demographic and clinical characteristics of the patient cohort used in the study, considering only moderate-to-severe steatosis.

Table S4 :
Carriers of the PNPLA3 I148M variant accounting for moderateto-severe steatosis.Lean patients have 18.5 kg/m 2 ≤BMI<25 kg/m 2 .Patients with steatosis have SHAD≥10 HU or LMA<40 HU.The rate of patients who are homozygous carriers by the chi-squared test: p=0.044 for lean patients with vs without steatosis, p<0.0001 for overweight patients with vs without steatosis, and 0.81 for lean vs overweight patients with steatosis.Supplementary TableS5: Demographic and clinical characteristics of the patient cohort used in the study, split into three BMI categories.
P-values were adjusted with the Benjamini-Hochberg method for a p<0.05 level of statistical significance.Lean patients have 18.5 kg/m 2 ≤BMI<25 kg/m 2 .Patients with steatosis have SHAD≥10 HU or LMA<40 HU.BMI=body mass index, SHAD=spleenhepatic attenuation difference, LMA=liver mean attenuation, CVD=cardiovascular disease, HTN=hypertension, T2DM=type 2 diabetes mellitus, HLD=hyperlipidemia.Supplementary P<0.0001 for all relevant rows.Age, BMI, LMA, and SHAD values listed for each group are in the format of median [interquartile range].A higher value for SHAD corresponds to increased hepatic steatosis.P-values were determined using the Kruskal-Wallis test for continuous fields (age) and chi-square for categorical fields (sex, race, CVD, HTN, index, SHAD=spleen-hepatic attenuation difference, LMA=liver mean attenuation, ALP=alkaline phosphatase, ALT=alanine transaminase, AST=aspartate transaminase.

Table S7 :
Uncorrected p-values from Dunn's test for multiple comparisons for biomarkers between the four groups of the patient cohort for mild to severe steatosis.

Table S8 :
Biomarker mean values and number of samples over the patient cohort, considering only moderate-to-severe steatosis.Supplementary TableS9: Z-scores and p-values from Dunn's test for multiple comparisons for biomarkers between the four groups of the patient cohort by the Kruskal-Wallis test, considering only moderate-to-severe steatosis.

Table S12 :
FIB-4 scores for lean, overweight, and obese patients with steatosis.P-values were calculated using the chi-square test and the Wilcoxon test in comparing proportion of patients with severe fibrosis and FIB-4 distributions, respectively.A FIB-4 score of FIB-4≥3.25 was considered to indicate severe fibrosis.Lean patients have 18.5 kg/m 2 ≤BMI<25 kg/m 2 , overweight patients have 25 kg/m 2 ≤BMI<30 kg/m 2 , and obese patients have BMI≥30 kg/m 2 .Patients with steatosis have SHAD≥-1 HU or LMA<40 HU.BMI=body mass index, SHAD=spleen-hepatic attenuation difference, LMA=liver mean attenuation.Supplementary TableS13: Top 200 phenotypes by significance of the association with SHAD in lean patients.Coefficients and p-values are from the SHAD x LEAN statistical interaction term of the logistic regression model controlling for age, sex, and race.LEAN is a binary characteristic indicating that a patient has 18.5 kg/m 2 ≤BMI<25 kg/m 2 .A p-value of p<7.52x10 -5 indicates significance with the Bonferroni multiple comparison correction.BMI=body mass index, SHAD=spleen-hepatic attenuation difference.Supplementary TableS14: Top 200 phenotypes by significance of the association with interaction between PDFF and lean characteristic in the UK Biobank cohort Flow chart illustrating the number of patients in the study cohort out of the total number with an available SHAD value after applying each exclusion criterion.Interfering conditions considered for this study were end-stage liver disease complications, alcohol use disorder, viral hepatitis, bariatric surgery, cachexia, and cancer.