For many patients with primary (e.g., hepatocellular carcinoma) and secondary liver tumors (e.g., metastases of gastrointestinal cancers), surgical tumor resection is not feasible due to advanced tumor stage or limited liver function at the time of tumor diagnosis [1
]. For these patients, transarterial chemoembolization (TACE) has evolved as a routinely performed treatment modality, providing local tumor control without major systemic side effects [3
]. However, response rates to TACE are heterogeneous and the appropriate selection of patients who will benefit particularly well from TACE in terms of overall survival is still debated [4
]. In the past, different single factors, such as the hepatic functional reserve, tumor distribution and size as well as laboratory parameters have been suggested for patient selection in the context of TACE [4
]. Moreover, predictive algorithms such as the assessment for retreatment (ART) score [6
] or SNACOR (tumour Size and Number, baseline Alpha-fetoprotein, Child-Pugh and Objective radiological Response)) score [8
], which are mainly based on imaging and on liver function, were suggested for guiding therapeutic decisions in these patients. However, since none of these parameters could be reliably validated as predictive and/or prognostic factors in larger clinical trials, patient allocation to TACE therapy is mainly still based on individual clinicians’ decision in daily routine, corroborating the vital need for novel pre-interventional stratification strategies.
Sarcopenia has been defined as the “progressive loss of muscle mass and strength with a risk of adverse outcomes such as disability, poor quality of life and death” by the Special Interest Group of the European Sarcopenia Working Group in 2010 [9
]. In patients with malignant diseases, muscle loss is induced by a systemic inflammatory response leading to decreased appetite, increased catabolism and immobility [10
]. Muscle loss was recently associated with an impaired prognosis in patients with different solid tumors [12
]. In patients with hepatocellular carcinoma (HCC) and colorectal cancer, sarcopenia has been associated with an impaired overall- and disease-free survival after surgical resection or radiofrequency ablation [13
]. However, the impact of sarcopenia on tumor response and overall survival in patients receiving TACE therapy in a palliative intent has not been sufficiently assessed to date. We therefore aimed at evaluating a potential role of sarcopenia as an easily accessible marker for predicting the clinical outcome of patients undergoing TACE for primary and secondary liver cancer.
2. Patients and Methods
2.1. Design of the Study and Patient Characteristics
This observational cohort study was conducted to analyze a potential impact of sarcopenia on treatment response as well as overall survival following TACE therapy in patients with primary (HCC) or secondary (metastases) liver cancer. 56 patients who received TACE therapy at the Department of Diagnostic and Interventional Radiology at University Hospital RWTH Aachen between 2013 and 2018 were subsequently enrolled into this study (see Table 1
for patient characteristics). The protocol of the study was approved by the ethics committee of the University Hospital RWTH Aachen (RWTH Aachen University, Germany) under the EK335/17. The study was conducted in accordance with the ethical standards laid down in the Declaration of Helsinki.
2.2. Transarterial Chemoembolization (TACE)
Hepatic tumors were treated with an emulsion of a chemotherapeutic agent and an embolic agent diluted with iodized contrast (Ultravist 300, Bayer Vital GmbH, Leverkusen, Germany). For HCCs, a combination of doxorubicin and ethiodized oil (Lipiodol, Guerbet LLC, Bloomington, IN, USA) was used. Liver metastases from colorectal, gastric or pancreatic cancer were treated using a chemotherapeutic agent in accordance with the respective guidelines as well as degradable starch microspheres (EmboCept S, PharmaCept GmbH, Berlin, Germany) or drug eluting beads (DcBeads, BTG International Ltd, London, UK). TACE procedures were conducted via the right femoral artery. A hepatography and a contrast-enhanced cone-beam computed tomography (CT) scan in late arterial contrast phase were performed using a distinct microcatheter. Depending on the tumor type and the number, size, localization and arterial supply of the tumor, a superselective (subsegmental), selective (segmental) or non-selective (lobar) approach was performed.
2.3. TACE Response Analysis
TACE response was analyzed on either a multidetector CT scan with multiphasic, contrast-enhanced acquisitions in unenhanced, arterial, portal venous and late-venous phase or a multiphasic, contrast enhanced liver magnetic resonance imaging (MRI) (1.5T, Philips Medical Systems DMC GmbH, Hamburg, Germany). Baseline imaging was conducted not earlier than 28 days before and at about four weeks after the TACE procedure. The median duration between the TACE procedure and the follow-up imaging scan for the assessment of tumor response was 34 days. To evaluate tumor response to TACE, CT and MRI scans were analyzed based on RECIST 1.1 criteria for non-arterially enhanced tumor entities (liver metastases) [15
] and mRECIST criteria for HCC [16
]. HCC tissue was differentiated from lipiodol deposition by comparison of unenhanced and arterially enhanced contrast scans. Overall tumor response was graded according to the RECIST 1.1/mRECIST standard nomenclature: Complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). Complete response and partial response were defined as an objective response (OR) [17
2.4. Assessment of Sarcopenia and Psoas Muscle Index
Sarcopenia was assessed by measuring the longest diameter (D1
) and the perpendicular diameter (D2
) of the right (ri) and left (le) psoas muscle on an axial CT scan. All diameters were measured in the same CT plane, which was usually between lumbar vertebral body (LVB) 3 and LVB 4. The individual course of sarcopenia after TACE therapy was assessed accordingly using the same CT plane. The median time to the follow-up scan after TACE therapy for the assessment of sarcopenia was 57.5 days (IQR: 43.75–73 days). An exemplary image of the psoas muscle measurement is displayed in Figure 1
. All measurements were then added and divided by four. To normalize the muscle mass for the patient’s height it was divided by the square of the patients’ height. The psoas muscle index (PMI) was consequently defined as following:
2.5. Measurement of Cytokine Serum Levels
Serum levels of interleukin (IL)-6 and IL-10 were analyzed using a multiplex immunoassay as described previously [18
]. Measurements were performed on a Bio-Plex 200 system using Bio-Plex Manager 6.0 software (Bio-Plex Pro Human Chemokine Panel, #171AK99MR2, Bio Rad, Hercules, CA, USA).
2.6. Measurement Laboratory Parameters
Standard laboratory markers were measured in the laboratory center for blood analyses at University Hospital RWTH Aachen. Standard haematological and clinical chemistry parameters were measured using the Sysmex XN9000 (Sysmex GmbH, Norderstedt, Germany) and the Cobas 8000 c701 (Hoffmann-La Roche AG, Basel, Switzerland).
2.7. Statistical Analysis
Statistical analyses were performed as recently described [19
]. In detail, Shapiro-Wilk test was performed to test for normal distribution of data. Non-parametric data for two groups were compared using the Mann-Whitney-U-Test. Receiver operating characteristic (ROC) curves were created by plotting sensitivity against 1-specificity. The predictive value of the PMI with respect to an objective response to TACE therapy was also evaluated using binary logistic regression. We plotted Kaplan-Meier curves to display the overall survival of patient subgroups with respect to different patient characteristics (e.g., PMI). The optimal cut-off value with respect to overall survival was calculated with a recently published biometric software, which fits Cox proportional hazard models to the survival status and the survival time [20
]. The prognostic relevance of individual variables was also tested by univariate and multivariate Cox-regression analyses. Parameters with p
-value < 0.2 in univariate analysis were included into multivariate analysis. The hazard ratio (HR) and the 95% confidence interval are displayed. Correlation analyses were performed using the Pearson correlation coefficient (r). All statistical analyses were performed with SPSS 23 (SPSS, Chicago, IL, USA) [21
]. A p
-value of <0.05 was considered statistically significant (* p
< 0.05; ** p
< 0.01; *** p
By using a well characterized cohort of 56 patients undergoing TACE therapy for primary or secondary liver cancer, we demonstrated that the presence of sarcopenia represents a negative prognostic parameter in terms of overall survival for these patients. In contrast, sarcopenia was not predictive for tumor response, highlighting that the influence of sarcopenia on the patients’ outcome did not reflect tumor specific factors but rather the patients’ general clinical condition.
In the last few years, multimodal therapies for primary and secondary hepatic malignancies have considerably evolved [23
] leading to a significantly improved outcome of patients with liver cancer [24
]. Beside highly intensive chemotherapies and radical surgical approaches, locally ablative techniques such as transarterial chemoembolization (TACE) have become of increasing importance in the therapy of both primary and secondary liver tumors [26
]. In this context, the question which individual patient represents an optimal candidate for different therapeutic modalities and, in particular, which patient will particularly well benefit, e.g., from TACE therapy, represents a highly debated but still unresolved clinical issue. In general, the individual therapeutic decisions vary significantly between centers, which are not only due to local preferences and experience but also due to a lack of well-established pre-interventional stratification tools. In most centers, the local tumor expansion (e.g., vascular invasion), liver function and clinical scores such as the hepatoma arterial embolization prognostic (HAP) score [29
] are considered when deciding whether a patient should receive TACE or not. Here, we show that a low pre-interventional psoas muscle index (PMI), as a surrogate for the presence of sarcopenia which can be calculated in routine pre-interventional staging CT or MRI scans, is indicative for the patients’ post-interventional overall survival (OS). As such, a pre-interventional PMI below our ideal cut-off value of 13.39 mm/m2
was an independent predictor for an unfavorable outcome (see Figure 3
and Table 2
). Patients with a PMI below this cut-off value showed a median OS of just 491 days (16.4 month) compared to 1291 days (43 month) in patients with a PMI above the ideal cut-off value. Off note, these results are in line with most available studies on an association between sarcopenia and survival in cancer, highlighting the value of the pre-intervention PMI in predicting patients’ prognosis after TACE therapy. Importantly, a combination of the PMI with other prognostically relevant factors such as serum albumin levels, further increased the PMI’s prognostic power arguing that this easily accessible score might also be implemented into existing and future stratification algorithms rather than being used as a stand-alone parameter. Nevertheless, this approach warrants further internal and external validation before a clinical implementation of, e.g., the PMI/albumin score into clinical routine can be considered.
Similar to the association between pre-interventional sarcopenia and a reduced OS, at a cut-off value of 11.48 mm/m2
, a low post-interventional PMI also turned out as a negative prognostic factor in our cohort of patients. We therefore attempted to answer the clinically more important question whether a longitudinal change of the PMI over time (between the pre- and post-interventional CT-scan) might provide further information on the patients’ outcome. Strikingly, this analysis demonstrated that patients whose PMI was decreasing after TACE had a significantly reduced overall survival compared to patients with an increasing PMI. Importantly, the difference in median OS (negative ∆PMI: 506 days (16.9 month) vs. positive ∆PMI: endpoint not reached yet) was even more apparent than that in patients with a high/low PMI at the pre-interventional time-point, highlighting that a progression of sarcopenia under TACE therapy represents a previously unrecognized prognostic factor in patients receiving TACE. Thus, we suggest that the PMI should not only be integrated into algorithms for the clinical decision making in patients eligible for TACE but might also be used to trigger specific measures in terms of nutritional support in patients with a low PMI. Notably, recent guidelines for the treatment of liver cancer have integrated the use of specific nutritional support strategies in the management of patients with HCC [30
]. Exemplary, the supplementation with branched-chain amino acids (BCAAs) and L-carnitine in patients with HCC was suggested to protect from deterioration of liver function and to increase muscle protein synthesis [31
]. Furthermore, different authors have demonstrated beneficial effects of physical exercise on muscle atrophy and weight loss in patients with HCC [33
]. However, further trials will be necessary to demonstrate that these measures are able to preserve skeletal muscle mass in patients receiving TACE and to improve the prognosis of these patients.
Despite the underlying pathophysiological mechanisms linking sarcopenia and especially a negative ∆PMI with an impaired patients’ prognosis is not fully understood, our observation that sarcopenia has no impact on the direct tumor response suggests that other mechanisms are involved in the underlying association between sarcopenia and an impaired patients’ survival. Sarcopenia and especially its end-stage in oncologic patients, the so-called cachexia, are associated with a systemic inflammatory response. In line, we found a significant positive correlation between the pre-interventional PMI and IL-10 serum levels as an example of an anti-inflammatory cytokine and a trend towards a positive correlation between the PMI and IL-6, supporting the association between systemic inflammation and sarcopenia. Therefore, it seems likely that an impaired nutritional state and/or a progressive deterioration of the nutritional state after TACE treatment reflects general adverse conditions (such as a systemic inflammatory response) that may have a considerable impact on patients’ out- comes upon intra-arterial treatments.
Although the present data suggested that PMI might be used to predict patients’ post-interventional outcome, they do not allow answering the question whether an individual patient with a low PMI might have benefitted similarly or even more in terms of long-term survival from a different treatment modality such as systemic treatment or other locally ablative therapies. Moreover, the number of patients included into this study is rather small as this study was designed as an exploratory analysis and therefore needs further validation. Finally, we included both patients with HCC and secondary liver cancer in our analysis. Although this aspect reduces the entity-specific validity of our findings and limits pathophysiological conclusion, it further suggests that the prognostic value of the PMI on patients’ outcome is rather treatment- than entity-related. To gain further insight into a potential entity-specific validity of the PMI, we conclusively performed all analyses in the subgroup of HCC patients alone (n
= 46, Figure S1
). These analyses revealed comparable results with respect to a prognostic relevance of the PMI and especially the PMI/albumin score following TACE therapy in HCC patients only, although statistical significance was not reached for the PMI only (most likely due to the reduced patient number). Again, a further post-interventional decrease of the PMI (negative ∆PMI) was a strong predictor of overall survival. To exclude a potential confounding effect of the patients’ liver function or alpha-fetoprotein (AFP) serum levels in the subgroup of HCC patients, we performed Cox-regression analysis which revealed no impact of bilirubin (HR: 0.634 (0.237–1.679), p
= 0.365) or AFP (HR: 1.000 (0.999–1.001), p
= 0.412) serum levels on patients’ overall survival in our cohort. Together, larger confirmatory prospective clinical studies including patients with different malignancies as well as alternative treatment modalities are warranted to exclude further potential confounders in multivariate Cox-regression analysis and to gain further information on a potential treatment predictive value of the PMI in patients with liver cancer, which we hope to have stimulated with this analysis.