Patient demographics

A total of 43 patients were registered in this study: 17 (39.5%) had cachexia, whereas cachexia was not observed in 26 (60.5%) patients. Cardiac 18F-FDG uptake was visually assessed based on the 5-PS scores [12, 13]. A previous study revealed that the optimal score for 18F-FDG uptake cut-offs for outcome in patients with NSCLC was 3 [13]. Therefore, a score of 3 was chosen as the cutoff point for further analysis, and all patients were divided into two groups; patients with scores of 1,2 or 3 were in the low group and those with scores of 4 or 5 were in the high group [13] (Fig. 1). Twenty-seven (62.8%) patients had high visual cardiac 18F-FDG uptake and 16 (37.2%) patients had low accumulation. The patient demographics according to visual cardiac 18F-FDG uptake are listed in Table 1. The median value of SUVmax by tumor and cardiac 18F-FDG uptake in 41 patients with stage IV was 9.5 (range, 3.2–21.7) and 4.3 (range, 1.6–11.8), respectively, indicating that the SUVmax of tumor 18F-FDG uptake was significantly higher than that of cardiac 18F-FDG uptake (p < 0.001). The median values of NLR, PLR, and PNI were 3.8 (range, 1.8–6.5), 213.5 (range, 117.4–295.1), and 43.9 (range, 40.7–63.0).

Fig. 1

18F-FDG PET findings before first-line therapeutic agent administration in patients with advanced NSCLC. PET imaging before any treatment showing representative images based on the definitions of the five-point scale (5-PS) scores using cardiac 18F-FDG uptake (red dotted frame, cardiac lesion; red arrow, liver). Score of 1: no uptake (A); score of 2: uptake lower than the mediastinum (B); score of 3: uptake higher than the mediastinum and lower than the liver (C); score of 4: uptake moderately higher than the liver (D); and score of 5: uptake markedly higher than the liver (E). Low cardiac visual scores are 1, 2, or 3, and high cardiac visual scores are 4 or 5. Abbreviations: 18F-FDG PET, two-deoxy-2-[fluorine-18]-fluoro-d-glucose positron emission tomography1; NSCLC, non-small cell lung cancer

Table 1 Patient’s demographics according to visual cardiac 18F-FDG uptake

Low visual cardiac 18F-FDG uptake was significantly observed in patients with non-AC (adenocarcinoma) (p = 0.030), cachexia (p = 0.004), high NLR (p = 0.016), high tumor MTV (p = 0.009), and high tumor TLG (p = 0.009). Table 2 presents the patient demographics according to cardiac 18F-FDG uptake by SUVmax. Patients with a low SUVmax for cardiac 18F-FDG exhibited a significantly higher NLR (p = 0.047), MTV (p = 0.009), and TLG (p = 0.009) than those with a high SUVmax. Regarding genetic alterations, EGFR mutation and anaplastic lymphoma kinase rearrangement were positive in nine and one patients, respectively. For first-line treatment, the nine patients harboring EGFR mutations received EGFR-TKIs (three patients in gefitinib and six patients in osimertinib), and one patient with anaplastic lymphoma kinase rearrangement was treated with alectinib. Thirty patients were treated with immune checkpoint inhibitors (ICIs), including 21 patients who received ipilimumab plus nivolumab, 5 patients received pembrolizumab, and 4 patients who received carboplatin, paclitaxel, bevacizumab and atezolizumab. Two patients received carboplatin plus nab-paclitaxel and one patient received cisplatin plus pemetrexed. The metastatic status of 41 patients with stage IV revealed that there are 10 patients with pulmonary metastasis, 12 patients with pleural metastasis, 6 patients with brain metastasis, 17 patients with bone metastasis, 5 patients with liver metastasis, 5 patients with adrenal metastasis, 12 patients with lymph node metastasis, 2 patients with skin metastasis, and one patient with other.

Table 2 Patient’s demographics according to cardiac 18F-FDG uptake

The findings of electrocardiogram showed that there are 8 patients with complete right bundle block, one patient with Movitz type II atrioventricular block, one patient with type 1 atrioventricular block, 4 patients with left ventricular hypertrophy, 2 patients with old myocardial infarction, one patient with left bundle block, one patient with supraventricular premature contractions, 2 patients with atrial fibrillation, one patient with premature ventricular contractions, and one patient with pacemaker. Regarding the information of diabetes mellitus, 7 patients with diabetes mellitus receive any medicine, and hemoglobin A1c of more than 6.5% was observed in 8 patients. The median value of blood sugar level before the performance of 18F-FDG PET imaging was 111.5 mg/dl, ranging from 88 to 179 mg/dl.

Relationship between 18F-FDG uptake and cachexia

The quantitative values of 18F-FDG uptake on PET were compared based on the presence or absence of cachexia (Fig. 2). No statistically significant differences in SUVmax (Fig. 2A) or SUVpeak (Fig. 2B) for 18F-FDG uptake were observed between patients with and without cachexia. However, the MTV (Fig. 2C) and TLG (Fig. 2D) for 18F-FDG uptake were significantly higher in patients with cachexia than in those without cachexia.The cardiac SUVmax (Fig. 2E) and SUVpeak (Fig. 2F) did not differ according to the presence of cachexia.

Fig. 2

Comparison of SUVmax (A), SUVpeak (B), MTV (C), TLG (D), cardiac SUVmax (E), and cardiac SUVpeak (F) according to cachexia presence. MTV and TLG were significantly higher in patients with cachexia than those without cachexia. Abbreviations: SUV, standardized uptake value; MTV, metabolic tumor volume; TLG, total lesion glycolysis

Correlation of 18F-FDG uptake with BMI or weight loss

Figure 3 presents the correlation among 18F-FDG uptake, BMI, and body weight loss. The amount of 18F-FDG accumulation based on SUVmax, SUVpeak, MTV, and TLG was not significantly correlated with BMI or weight loss. Although there was no close correlation between weight loss and cardiac SUVmax or SUVpeak, the amount of 18F-FDG uptake according to the cardiac SUVmax was significantly correlated with BMI (Fig. 3E).

Fig. 3

Correlation of 18F-FDG uptake with different variables. Pearson’s correlations of BMI with SUVmax (A), SUVpeak (B), MTV (C), TLG (D), cardiac SUVmax (E), and cardiac SUVpeak (F) were performed. SUVmax (G), SUVpeak (H), MTV (I), TLG (J), cardiac SUVmax (K), and cardiac SUVpeak (L) correlated with weight loss (%).Abbreviations: 18F-FDG, Two-deoxy-2-[fluorine-18]-fluoro-d-glucose; BMI, body mass index; SUV, standardized uptake value; MTV, metabolic tumor volume; TLG, total lesion glycolysis; 95% CI, 95% confidence interval

Relationship between visual cardiac 18F-FDG uptake and different variables

This study examined the close association between visual cardiac 18F-FDG uptake and different variables such as white blood cell count, neutrophil, lymphocyte, platelet, total protein, albumin, lactate dehydrogenase, blood urea nitrogen, creatinine, CRP, GPS, NLR, PLR, and PNI. High BMI (Fig. 4A), low weight loss (Fig. 4B), low platelet count (Fig. 4C), and low CRP levels (Fig. 4D) were significantly associated with high visual cardiac 18F-FDG uptake. Besides, the cardiac SUVmax was not significantly correlated with tumor SUV [r = -0.144, 95% confidence interval (CI) -0.429 to 0.166, p = 0.361], MTV (r = -0.219, 95% CI -0.504 to 0.107, p = 0.181), and TLG (r = -0.290, 95% CI -0.558 to 0.032, p = 0.077). There was not significant correlation between cardiac SUVmax and NLR (r = -0.005, 95% CI -0.301 to 0.299, p = 0.997).

Fig. 4

Comparison of BMI (A), weight loss (%) (B), Plt (C), and CRP (D) according to cardiac visual score on 18F-FDG uptake. Abbreviations: 18F-FDG, Two-deoxy-2-[fluorine-18]-fluoro-d-glucose; BMI, body mass index; Plt, platelet; CRP, C-reactive protein

Survival analysis according to cardiac visual score and cachexia

The follow-up period was 549 (range, 43–1049) days. The median PFS and OS after the initial treatment were 208 and 549 days, respectively. Thirty-three patients experienced disease progression, and 25 died due to the primary disease. The Kaplan–Meier survival curve according to the cardiac visual score and cachexia is presented in Fig. 5. No statistically significant differences in PFS [median survival time (MST), 180 days vs. 238 days] (Fig. 5A) or OS (MST, 418 days vs. 716 days) (Fig. 5B) were observed between patients with and without cachexia. Although there was no significant difference in PFS (MST, 269 days vs. 164 days) (Fig. 5C) between patients with high and low cardiac visual scores, a significant difference in OS (MST, not reached vs. 362 days) (Fig. 5D) was observed between the two groups.

Fig. 5

Kaplan–Meier curves of PFS and OS according to cachexia and cardiac visual scores. No statistically significant difference for the PFS (A) and OS (B) was observed between the patients with and without cachexia. There was no significant difference for PFS between a cardiac high and low visual score (C); however, the patients with a low visual score yielded a worse OS than those with a high score (D). The absence of cachexia and a cardiac high visual score were not identified as a prognostic predictor of PFS (E) and OS (F), whereas, cachexia and a cardiac low visual score depicted a significant predictor for OS (H) but not PFS (G). Abbreviations: PFS, progression-free survival; OS overall survival

The absences of cachexia with a high cardiac visual score was not identified as a prognostic predictor for PFS (MST, 269 days vs. 179 days) (Fig. 5E) and OS (MST, 862 days vs. 733 days) (Fig. 5F), whereas cachexia and a low cardiac visual score was a significant predictor for OS (MST, 307 days vs. 950 days) (Fig. 5H) but not PFS (MST, 178 days vs. 238 days) (Fig. 5G).

Next, univariate analysis of all patients identified histology, NLR, GPS, cardiac visual score, cardiac SUVmax, and, tumor TLG, as significant predictors of OS (Table 3). The univariate log-rank test enabled screening for variables with p < 0.05 for subsequent multivariate analysis. As a significant predictor for PFS was not observed, we examined the multivariate analysis for predictors of OS. Multivariate analysis identified NLR and cardiac SUVmax as independent predictors of OS. Moreover, the cut off values for cardiac SUVmax and NLR were examined by receiver operating characteristic curve (ROC) analysis and sensitivity and specificity were calculated to determine the optimal cut-off value for differentiating cachexia ( +) from cachexia (-) using ROC curves. As a result, the optimal cutoff values for the cardiac SUVmax and NLR as determined by ROC curves analyses were 4.2 (sensitivity: 58.8%, specificity: 57.7%) and 3.9 (sensitivity: 64.7%, specificity: 69.3%), and the areas under the curve in the ROC analysis were 0.563 (cardiac SUVmax) and 0.660 (NLR). These cut-off values were almost similar to those of median cut-off value. Therefore, the results of survival analyses for the cardiac SUVmax and NLR were corresponding to those for Table 3.

Table 3 Univariate and multivariate survival analysis