Shallow and reverse attention network for colon polyp segmentation.

Polyp segmentation is challenging because the boundary between polyps and mucosa is ambiguous. Several models have considered the use of attention mechanisms to solve this problem. However, these models use only finite information obtained from a single type of attention. We propose a new dual-attention network based on shallow and reverse attention modules for colon polyps segmentation called SRaNet. The shallow attention mechanism removes background noise while emphasizing the locality by focusing on the foreground. In contrast, reverse attention helps distinguish the boundary between polyps and mucous membranes more clearly by focusing on the background. The two attention mechanisms are adaptively fused using a "Softmax Gate". Combining the two types of attention enables the model to capture complementary foreground and boundary features. Therefore, the proposed model predicts the boundaries of polyps more accurately than other models. We present the results of extensive experiments on polyp benchmarks to show that the proposed method outperforms existing models on both seen and unseen data. Furthermore, the results show that the proposed dual attention module increases the explainability of the model.

The absence of coronary artery calcification does not rule out the presence of significant coronary artery disease in Asian patients with acute chest pain.

The absence of coronary artery calcification (CAC) has been used to as an indication to rule out significant coronary artery disease (CAD). However, diagnostic usefulness of 'zero calcium score criteria' as a decision-making strategy to rule out significant CAD as the etiology of acute chest pain has not been studied in depth, especially in Asian ethnicity. We prospectively enrolled 136 Korean patients (58% men, 56 ± 13 years) who presented to the emergency department (ED) with acute chest pain and non-diagnostic ECG. All patients underwent 64-slice CT for calcium scoring and coronary CT angiography (cCTA). We investigated the association of CAC with the presence of ≥50% CAD on cCTA and with a final diagnosis of an acute coronary syndrome (ACS). Ninety-two patients out of 136 (68%) did not show detectable CAC, and 14 out of these 92 without CAC (15%) had ≥50% CAD on cCTA. Sensitivity, specificity, positive predictive value and negative predictive value of zero calcium score criteria for the detection of ≥50% CAD were 0.66 (95% confidence interval, 0.50-0.80), 0.83 (0.74-0.90), 0.64 (0.48-0.77), 0.85 (0.75-0.91), respectively. Patients who had ≥50% CAD without detectable CAC were younger (P = 0.001), and had a higher prevalence of smoking (P = 0.048) as compared to patients with a degree of CAC. Most of the patients with ≥50% CAD of non-calcified plaque were younger than 60 years of age (79%, 11/14), however, 3 of them were older than 60 years of age. Forty-five patients (33%) were subsequently diagnosed as having ACS, and 38% (17/45) of them had no CAC. Zero calcium score did not necessarily guarantee the absence of significant CAD, even in patients older than 60 years, in Asian ethnicity presenting to the ED with chest pain.