Our study aimed to evaluate whether quantitative imaging parameters obtained from a preoperative CT-scan could be used to predict mortality in patients with suspected AMI.

According to the literature the imaging finding with prognostic relevance is pneumatosis intestinalis with a reported odds ratio of 2.86 for short-term mortality [11]. In another study investigating prognostic CT findings, only the concomitant presence of pneumatosis and porto-mesenteric venous gas showed an odds ratio of 1.95 (95%CI 0.49–7.79, p = 0.34) for the occurrence of transmural necrosis [12]. Yet, other CT findings such as mural thickening, bowel distension, pneumatosis alone, and presence of thrombi or emboli were not different in patients with and without transmural necrosis [12].

This highlights the importance of new CT imaging markers to stratify patients at risk.

Another recent study investigated the prognostic relevance of CT-defined body composition parameters in patients with AMI to extract novel prognostic markers from the CT images [13]. This is especially of importance as even established laboratory parameters comprising lactate, D-dimers, C-reactive protein, and white blood cell count have low accuracy to predict a poor clinical outcome in these patients [14].

That is why there is definite need for further quantitative parameters to correctly predict a poor clinical outcome in patients with AMI. This would be crucial as the investigated radiodensities parameters in the current study were already demonstrated to be of prognostic relevance in critically ill patients [5,6,7].

In a patient collective with various underlying pathologies, Winzer et al. found the adrenal-to-spleen ratio of a portal venous scan to be a good predictor for 72-h mortality with an AUC of 0.94 using a cut-off value of > 1.4 [5]. This resulted in a specificity above 98% and a sensitivity at around 80% in a patient cohort of 203 patients [5]. The authors included different underlying causes of critical illness, most commonly with postoperative complications (128 patients) and septic shock (57 patients). Notably, the mortality was significantly lower in the patient cohort by Winzer et al. with 9.9% after 24 h and 24.1% after 72 h compared with the present study [5]. This might explain the divergent results regarding the accuracies.

In the present patient collective of patients with AMI, we found significant difference in means of the adrenal-to-spleen ratio especially regarding 24 h- and 30-day mortality. However, the diagnostic accuracy remained low with a maximum of 0.70 for 24-h mortality. Using the same cut-off value of > 1.41 as proposed by Winzer et al. [5] only a sensitivity of 34.6% and specificity of 86.5% could be achieved. Even with a more moderate cut-off value of > 1.103 achievable sensitivity and specificity remained poor with 61.5% and 73.1%, respectively.

Presumably, the identified differences of the investigated CT radiodensities could be explained by differences in the clinical conditions and the shock state of the investigated patient cohorts. The underlying causes of critical illness were more heterogeneous in the patient sample of Winzer et al. [5].

Hyperattenuation of the adrenal glands and hypoperfusion (and therefore hypoattenuation) of the spleen are part of the CT hypoperfusion complex which consists of a multitude of signs related to shock and shock-related hypotension and were reported to be of diagnostic and prognostic importance [9, 15,16,17].

Since the patients in our collective received the CT-scan in an acute setting, patients might have been less critically ill, and the scan was possibly performed shorter after the onset of symptoms and therefore less likely to have developed shock and sufficient changes in attenuation of the adrenal glands and spleen compared to the patient cohort of Winzer et al.

In contrast to the above-mentioned studies, one study found the absence of hyperattenuation of the adrenal glands to be related to mortality over 28 days for critically ill patients in septic shock [8]. This might indicate that also less contrast enhancement of the adrenal glands is in some cases a sign for worse prognosis.

Another reason for the identified results might be that the perfusion of the adrenal glands is reduced due to vessel stenosis, a common precondition of patients with AMI. In the study by Winzer et al. patients with stenosis of the celiac trunk were excluded from the analysis [5].

The prognostic relevance of adrenal gland enhancement was also evaluated in other diseases. Schek et al. [18] investigated 292 polytrauma patients in a retrospective analysis. Of these, 18 patients (6.1%) showed a stronger enhancement of the adrenal gland compared to the IVC, which was a statistically significant predictor for poor clinical outcome. Moreover, mean adrenal enhancement was significantly higher in non-survivors compared with survivors (101.9 ± 40.6 HU vs. 86.1 ± 27.0 HU; p < 0.001).

Boos et al. investigated 88 patients of the ICU with clinical deterioration. 43.2% of these patients showed a high adrenal gland enhancement, which was also a predictor for poor outcome. Of this patient cohort, overall, 50% died, whereas only 16% died of the patient group with regular adrenal gland enhancement [19].

In another study investigating 71 adult patients with hypovolemic shock found a statistically significant reduction, not only of splenic volumes, but also in splenic attenuation values compared with a control population (105 ± 34 HU vs. 134 ± 25 HU; p < 0.001) [20]. Despite these promising results of the prognostic relevance of adrenal gland enhancement, there is still need for further investigation for clinical translation.

The present study has limitations. First, it is a retrospective single-center study with known inherent bias. Second, despite the quantitative nature of the measurements, there can still be some interreader variability. However, in our analysis, the interreader variability was low, which is consistent with a previous study [7]. Third, there might be some dependency of the adrenal gland enhancement on contrast media type, concentration, and flow, which reduces possible external validity. Forth, it is important to discuss the selection of only surgical patients in the present analysis. The generalization of our results to all patients with AMI remains limited, as it cannot be adjusted for patients with less severity, which could be treated conservatively. Fifth, co-morbidities, such as congestive heart failure or sepsis, could have a confounding role on the investigated radiodensities. Due to the study design, there was not enough data to adjust for these co-morbidities. There is also the potential heterogeneity of contrast enhancement of the spleen especially in earlier contrast media phases.

In conclusion, the contrast media enhancement of the adrenal gland is associated with the 24-h and 30-day mortality in patients with AMI. However, the prognostic relevance needs to be validated in other cohorts for a possible translation into clinical routine.