Accuracy of controlled attenuation parameter for assessing liver steatosis in individuals with morbid obesity before bariatric surgery

The ultrasound‐based controlled attenuation parameter (CAP) is a non‐invasive tool widely validated for assessing liver steatosis across different etiologies. However, few studies, with liver biopsy available, have investigated its performance in individuals with morbid obesity. Herein, we aimed to evaluate the diagnostic accuracy of CAP in participants with morbid obesity from the MAFALDA study before bariatric surgery.


| INTRODUC TI ON
Non-alcoholic fatty liver disease (NAFLD) is becoming the major cause of chronic liver disease worldwide, paralleling the epidemics of obesity and type 2 diabetes. The prevalence of NAFLD in the general population is estimated to be approximately 25% but it reaches over 90% among individuals with severe obesity undergoing weight loss procedures. 1 NAFLD/non-alcoholic steatohepatitis (NASH) are serious obesity-related complications and their presence qualifies individuals to be eligible for bariatric surgery. 2 Therefore, an accurate diagnosis and staging of NAFLD play an important role in the clinical management of obesity and decision-making on the most appropriate treatment modality (ie, lifestyle changes, pharmacological therapy, or bariatric surgery).
Although the stage of liver fibrosis is the main determinant of liver-related outcomes, 3 the degree of liver fat accumulation has a crucial role in the physiopathology of NAFLD and is causally related to liver damage. 4,5 Consistently, novel drugs under clinical investigation for the treatment of NAFLD primarily target the reduction of liver steatosis. 6 To date, liver biopsy represents the most accurate technique for the assessment of hepatic steatosis, necroinflammation, and fibrosis. 7 However, this procedure is considered invasive and accompanied by a low, but not negligible, risk of complications, including hepatic bleeding, perforation of adjacent organs, and infection. 8 Therefore, in the last years, an intensive effort has been carried out to identify reliable non-invasive diagnostic tests for NAFLD/NASH detection and to follow its course over time.
The ultrasound-based controlled attenuation parameter (CAP) is a non-invasive tool for the assessment of liver steatosis. CAP has primarily been validated among individuals with chronic liver disease due to different etiologies, including NAFLD and viral hepatitis. [9][10][11][12] To reduce the failure rate of transient elastography with the M probe due to high body mass index (BMI), the XL probe has been recently developed and specifically designed for individuals with overweight/obesity (skin-to-liver capsular distance ≥25 mm). [13][14][15] Several studies have examined the feasibility of XL-CAP for liver steatosis among individuals with overweight/obesity. 12,[15][16][17][18][19][20] However, a handful of histologically characterized cohort studies have specifically investigated the use of CAP for assessing liver steatosis in individuals with morbid obesity. [21][22][23] Via Alvaro del Portillo 200, 00128 Rome, Italy. Email: u.vespasiani@unicampus.it The aim of this study was to evaluate the diagnostic accuracy of CAP measured by the XL probe to detect liver steatosis in participants with morbid obesity from MAFALDA, a study in which liver biopsy was available. Additionally, we assessed the performance of FibroScan-AST (FAST) score 24 for detecting progressive NASH (NASH + NAFLD activity score [NAS] ≥ 4 + F ≥ F2) in this cohort.

| MAFALDA cohort
The "Molecular Architecture of Fatty Liver Disease in patients with obesity undergoing bAriatric surgery (MAFALDA)" study aims to understand the genetic and molecular architecture of NAFLD in individuals with morbid obesity. The MAFALDA study is composed by two substudies: MAFALDA-1 and MAFALDA-2.
MAFALDA-1 started on 22nd May 2020 and ended on 22nd June 2021. A total of 318 individuals were enrolled and plasma, serum, urine, and blood in EDTA were collected. Among the 318 participants, a total of 264 individuals underwent liver and visceral adipose tissue biopsy and all these individuals had a diagnosis of NAFLD assessed by histology from two pathologists from the Pathology Unit at "Campus Bio-Medico University Hospital" in Rome, Italy.
MAFALDA-2 started on 19th August 2021 and as for the 31st of October 2021 comprises 52 participants and it is still ongoing.
In MAFALDA-2 plasma, serum, urine, and blood in EDTA were and will be collected. In this study, NAFLD diagnosis is and will be assessed by vibration-controlled transient elastography including CAP measurement. The MAFALDA study has been approved by the Local Research Ethics Committee (no. 16/20) and it was conducted in accordance with the principles of the Declaration of Helsinki. All participants gave written informed consent to the study.
At the preoperative assessment visit, the following clinical and anthropometric data were recorded: age, gender, BMI, waist and hip circumferences, systolic and diastolic blood pressure, medical history, current pharmacological therapy, lifestyle, and family history of major diseases. BMI was calculated by dividing the weight (kg) by the square of the height (m 2 ). Waist and hip circumferences were measured at the umbilicus level and around the widest portion of the buttocks, respectively, using non-stretchable tape.
Additional blood samples and first morning midstream urine samples were also collected. The different blood fractions (plasma, serum, buffy coat) were separated by centrifugation at 3000 g for 10 min at room temperature.
Type 2 diabetes was diagnosed according to the American Diabetes Association criteria. 28 Hypertension was defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg (measured on two different days), or treatment with anti-hypertensive drugs. Dyslipidemia was defined as total cholesterol ≥200 mg/dL and/or triglycerides ≥150 mg/dL or treatment with lipid-lowering drugs. Metabolic syndrome was diagnosed according to the International Diabetes Federation criteria. 29 Insulin resistance (IR) was assessed by the homeostatic model assessment for insulin resistance (HOMA-IR) and defined by a cut-off of ≥2.5. 30 A hypocaloric diet (1065 kcal, 39% carbohydrates, 25% fats, 38% proteins) was recommended two weeks before the surgery by a dietician.

| Collection of tissue biological samples and histopathological evaluation
On the day of surgery, a laparoscopic-guided percutaneous liver core biopsy was obtained by experienced bariatric surgeons, using a 16 gauge-20 mm modified Menghini needle with automatic aspiration (Biomol, HS Hospital Service, Rome, Italy). The liver biopsy was performed during the first part of the surgical procedure, within 10 min after pneumoperitoneum installation, in the left lobe of the liver.
Adequacy of the liver biopsy was assessed macroscopically, and an additional core was collected if the specimen was less than 8 mm in length. The biopsy specimen was fixed in formalin, embedded in paraffin, and stained with hematoxylin-eosin and picrosirius red for fibrosis evaluation. The slides were analysed by two experienced pathologists (GP and SC) blinded to clinical and laboratory data.
Liver biopsy specimens included in histological analysis were at least 8 mm in length and contained at least 7 portal tracts.
Steatosis grade (from 0 to 3), lobular inflammation grade (from 0 to 3), ballooning grade (from 0 to 2), and fibrosis stage (from 0 to 4) were scored according to NAFLD activity score (NAS) classification. 31 NASH diagnosis was established by the pathological assessment based on Brunt et al criteria. 32 If the biopsy specimen exceeded the minimum length required for histological evaluation, part of the exceeding specimen was immediately snap frozen in liquid nitrogen. Additionally, a wedge visceral fat biopsy was performed at gastric cardia and immediately snap frozen in liquid nitrogen.

| Vibration-controlled transient elastography including CAP measurement
CAP was acquired simultaneously with liver stiffness measurement (LSM) using the vibration-controlled transient elastography device FibroScan ® (Echosens, Paris, France) within three months before the surgery by an experienced operator blinded to clinical and laboratory data, according to the manufacturer's instructions. All individuals were fasted at least three hours before the examination and were placed in the supine position with their right arm fully abducted.
Measurements were performed on the right lobe of the liver through an intercostal space using the XL probe. 13 The success rate was calculated as the number of successful measurements divided by the total number of measurements obtained. Only examinations with at least 10 successful acquisitions, with an interquartile range/median (IQR/M) of LSM ≤30%, and with a success rate ≥60% were deemed valid. 33

| Statistical analyses
All analyses were performed using the software R, version 4.0.4. 34

Continuous variables were shown as mean (standard deviation [SD])
or median (IQR), as appropriate. Categorical variables were shown as number (percentage). Differences between CAP and steatosis grades were assessed using Kruskal-Wallis test.
The overall accuracy of CAP and FAST score for detecting liver steatosis and progressive NASH, respectively, was estimated by the area under the receiver-operating characteristics curve (AUROC) using the pROC package. 35 The optimal cut-off values of CAP for S ≥ S1, S ≥ S2, and S = S3 disease were chosen at points with the highest Youden index. The population was randomly selected using a bootstrap method, repeating the process 10,000 times.
The impact of covariates on CAP values was assessed by a multivariable linear regression model. All statistical tests were two-sided. P values <0.05 were considered statistically significant.  Table 1. The mean (SD) age and BMI were 43.8 (10.1) years and 41.0 (4.0) kg/m 2 , respectively. More than two out of three individuals were women. The prevalence of type 2 diabetes was 19%, whereas those of hypertension and metabolic syndrome were 47% and 72%, respectively. Thirty-five (29%) subjects did not have steatosis (S0), whereas NASH was diagnosed in 54 (45%) individuals. Sixty-four (53%) individuals underwent FibroScan ® examination within two weeks before the surgery. A total of 40 (33%) individuals followed the hypocaloric diet recommended before surgery for more than one week. Individuals with higher steatosis have higher transaminases, fasting glucose, HbA1c, triglycerides, and insulin resistance compared to those with lower steatosis. Moreover, these individuals had higher rates of hypertension, metabolic syndrome, and significant fibrosis. Details of liver histology of the cohort according to NAS scoring system are provided in Supplementary Table S1.
We also examined the performance of the recently published

| Accuracy of FAST score
Finally, we assessed the performance of the recently designed FibroScan-AST (FAST) score for detecting NASH with significant activity and fibrosis. 24 We found that the diagnostic accuracy of FAST score for detecting progressive NASH (NASH + NAS ≥ 4 + F ≥ F2) as assessed by AUROC was 0.76 (95% CI 0.66-0.86) ( Table 4).  we found that the performance of CAP for grading liver steatosis using the XL probe was excellent (AUROCs 0.91, 0.83, and 0.86 for S ≥ S1, S ≥ S2, and S = S3, respectively). The CAP cut-offs of 300, 328, and 344 dB/m had 82%, 87%, and 94% sensitivity for S ≥ S1, S ≥ S2, and S = S3, respectively. The diagnostic accuracy of CAP using the XL probe for predicting steatosis in our cohort was consistent with and even higher than that reported in previous studies including individuals with morbid obesity. [21][22][23] As an example, in a prospective cohort study of 76 individuals, Garg   On the other hand, the diagnostic performance of CAP was better in individuals undergoing bariatric surgery (AUROCs of 0.80, 0.77, and 0.76 for S ≥ S1, S ≥ S2, and S = S3, respectively). As commented by the authors, "This unexpectedly increased performance of CAP in bariatric cohorts may be due to the highly standardized recruitment and liver biopsy procedure, irrespectively of liver disease severity.
Consequently, studies with bariatric cohorts are usually characterized by a control group who does not have steatosis, resulting in fewer grey zone subjects with S1 and S2 and in a better diagnostic performance of CAP in these cohorts." 19 Consistently, in our cohort, 35 (29%) participants did not have hepatic steatosis. Notably, this rate was higher than in other bariatric cohorts in which ranged from 8% to 19%. [21][22][23] Larger histologically characterized cohort studies in this high-risk population are needed to clarify these findings.
Notably, our CAP cut-offs closely resembled those reported by Garg et al 21 and by Naveau et al 22 (Supplementary Table S3).
Additionally, they were similar to CAP-XL cut-offs recently identified by Petroff et al 19 which showed good sensitivity, but lower specificity compared with ours ( The main strength of our study is the use of standardized procedures for clinical data and biological samples collecting, processing, and storage. Furthermore, the mean biopsy length of our cohort was higher than 1.5 cm. Limitations include (1) the small sample size of the cohort, (2) the lack of an external validation, and (3) the inability to test the performance of LSM for advanced fibrosis due to the low prevalence of F3-F4 fibrosis in our cohort (only three subjects, 2%).
However, we validated the model internally using 10,000 bootstrap samples and we provided a confidence interval for the optimal cutoffs unlike the standard ROC methods. The performance of LSM for significant fibrosis (F ≥ F2) and the impact of covariates on LSM values are described in supplementary material (Supplementary Tables   S4 and S5).
To conclude, CAP measured by the XL probe represents a powerful non-invasive tool for grading liver steatosis in individuals suffering from morbid obesity. Assessing liver steatosis by CAP may help detect hepatic steatosis in individuals with morbid obesity, thus rendering them eligible for bariatric procedures, and may help the follow-up of this disease after bariatric surgery.

ACK N OWLED G EM ENTS
We thank Dr Chiara Magistri, Dr Roberta Veralli, and all nurses and receptionists at Campus Bio-Medico University Hospital for their hard work and essential support in this project.

CO N FLI C T S O F I NTE R E S T
The authors declare no financial or other relationships with drug manufacturers that could lead to a conflict of interest.

AUTH O R CO NTR I B UTI O N S
UVG and SR contributed to study concept and design. FT, UVG, and SR contributed to drafting the manuscript. FT and ADV contributed to perform the statistical analysis. All authors contributed to analysis and interpretation of data, critically revised the manuscript for important intellectual content and approved the final version for submission. All authors had fully access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.