Study Design and Patient Selection

This was a retrospective observational study involving two distinct patient cohorts:

• mRCC Cohort: 29 patients diagnosed with metastatic renal cell carcinoma, treated with a combination of Ipilimumab and Nivolumab.
• mNSCLC Cohort: 37 patients diagnosed with metastatic non-small cell lung cancer, treated with Pembrolizumab as a first-line therapy.

Patients were selected based on the availability of pre-treatment and follow-up imaging data and inflammatory marker measurements. Clinical and laboratory data were collected at baseline and three months after initiation of immunotherapy.

Body Composition Analysis

Axial CT scans reconstructions at the L3 vertebra level were used to evaluate body composition parameters. Image analysis was performed using a semi-automated segmentation software. Specific measurements included:

• Skeletal Muscle Area (SMA) – measured by segmenting rectus abdominis, psoas, and paravertebral muscles.
• Visceral Fat Area (VFA) and Subcutaneous Fat Area (SFA) – assessed separately and normalized for height, obtaining Visceral Fat Index (VFI) and Subcutaneous Fat Index (SFI)
• Adjusted Skeletal Muscle Index (aSMI) – a refined index considering skeletal muscle density and tissue quality.

Calculation of aSMI

The accurate measurement of muscle area is often complicated by the presence of interspersed adipose tissue, which can alter surface area estimations depending on whether it is included in the segmentation. Traditional methods sometimes neglect this fat fraction, leading to an overestimation of true muscle area. To improve precision, a commonly used technique employs an approach based on Hounsfield Unit (HU) thresholds to distinguish muscle from fat. However, HU-based segmentation presents challenges, as baseline attenuation values can be influenced by factors such as contrast media administration, scanner specifications, and imaging noise.

In accordance with best practices, our institution adheres to the ALARA (As Low As Reasonably Achievable) principle; therefore, staging and re-staging CT scans do not routinely include non-contrast acquisitions, making HU-based corrections unreliable. Additionally, variations in tube voltage (kVp) and milliamperage modulation can further affect attenuation values, complicating direct tissue segmentation.

To overcome these limitations, we implemented an alternative method based on a compositional ROI approach. This involved:

1. Defining a comprehensive ROI: Encompassing all skeletal muscle in the L3 slice, along with a portion of interspersed adipose tissue; other than this ROI we estimated the area (Atot) and the global attenuation (Dtot).
2. Measuring reference densities: Independent ROIs were placed over purely skeletal muscle (Dta) and fat regions (Dtm) in the same CT slice to establish their respective HU values avoiding bias due to contrast media administration, tube voltage and amperage modulation.
3. Estimating tissue composition: The proportion of muscle and fat within the primary ROI was computed based on known reference HU values (see formula in fig.1)
4. Final computation of aSMI: The corrected SMA area, reflecting only muscle mass, was normalized by height squared to obtain aSMI.

Fig 1: Method of measurement of aSMI form SMI and densities of pure adipose tissue and muscolar tissue in the axial image; this value, alternative of the use of HU thresholds for differentiating adipose tissue and muscle, is independent from the administration of IV contrast media, mAh and kV used in the exam because of the internal standardisation of the image.

This methodology allowed for an accurate and reproducible quantification of muscle mass, minimizing the confounding effects of adipose infiltration. The aSMI was calculated alongside traditional SMI values in all patients, providing a more refined prognostic tool.

Inflammatory Markers

Inflammation-based indices were computed from routine blood tests, including:

• Neutrophil-to-lymphocyte ratio (NLR) – a well-established marker of systemic inflammation and immune status.
• Platelet-to-lymphocyte ratio (PLR) – indicative of systemic thrombosis and inflammation.
• Monocyte-to-lymphocyte ratio (MLR) – associated with immune suppression and tumor progression.
• Systemic Immune-Inflammation Index (SII) – calculated as (neutrophils × platelets) / lymphocytes, which has been linked to worse prognosis in multiple cancer types.

Statistical Analysis

Group differences were analyzed using the Mann-Whitney U test, while correlations between body composition parameters, inflammatory indices, and survival outcomes were evaluated through Cox proportional hazards regression. Kaplan-Meier survival analysis was employed to estimate overall survival (OS) differences based on body composition and inflammatory markers.