LN-Net: Perfusion Pattern-Guided Deep Learning for Lymph Node Metastasis Diagnosis Based on Contrast-Enhanced Ultrasound Videos

OBJECTIVE

The blood flow in lymph nodes reflects important pathological features. However, most intelligent diagnosis based on contrast-enhanced ultrasound (CEUS) video focuses only on CEUS images, ignoring the process of extracting blood flow information. In the work described here, a parametric imaging method for describing blood perfusion pattern was proposed and a multimodal network (LN-Net) to predict lymph node metastasis was designed.

METHODS

First, the commercially available artificial intelligence object detection model YOLOv5 was improved to detect the lymph node region. Then the correlation and inflection point matching algorithms were combined to calculate the parameters of the perfusion pattern. Finally, the Inception-V3 architecture was used to extract the image features of each modality, with the blood perfusion pattern taken as the guiding factor in fusing the features with CEUS by sub-network weighting.

DISCUSSION

The average precision of the improved YOLOv5s algorithm compared with baseline was improved by 5.8%. LN-Net predicted lymph node metastasis with 84.9% accuracy, 83.7% precision and 80.3% recall. Compared with the model without blood flow feature guidance, accuracy was improved by 2.6%. The intelligent diagnosis method has good clinical interpretability.

CONCLUSION

A static parametric imaging map could describe a dynamic blood flow perfusion pattern, and as a guiding factor, it could improve the classification ability of the model with respect to lymph node metastasis.

Influence of traumatic brain injury on ipsilateral and contralateral cortical perfusion in mice

Traumatic brain injury (TBI) is one of the most important causes of death in young adults. After brain injury cortical perfusion is impaired by cortical spreading depression, cerebral microvasospasm or microvascular thrombosis and contributes to secondary expansion of lesion into surrounding healthy brain tissue. The present study was designed to determine the regional cortical perfusion pattern after experimental TBI induced by controlled cortical impact (CCI) in male C57/BL6N mice. We performed a longitudinal time series analysis by Laser speckle contrast imaging (LSCI). Measurements were carried out before, immediately and 24 h after trauma. Immediately after CCI cortical perfusion in the lesion core dropped to 10 % of before injury (baseline; %BL) and to 21-24 %BL in the cortical area surrounding the core. Interestingly, cortical perfusion was also significantly reduced in the contralateral non-injured hemisphere (41-58 %BL) matching the corresponding brain region of the injured hemisphere. 24 h after CCI perfusion of the contralateral hemisphere returned to baseline level in the area corresponding to the lesion core, whereas the lateral area of the parietal cortex was hyperperfused (125 %BL). The lesion core region itself remained severely hypoperfused (18 to 26 %BL) during the observation period. TBI causes a maldistribution of both ipsi- and contralateral cerebral perfusion immediately after trauma, which persist for at least 24 h. Higher perfusion levels in the lesion core 24 h after trauma were associated with increased tissue damage, which supports the role of reperfusion injury for secondary brain damage after TBI.

Perfusion analysis of benign parotid gland tumors by contrast-enhanced ultrasonography (CEUS)

BACKGROUND

Diagnosis of parotid gland tumors is sometimes challenging due to their diversity and pleomorphic histological appearance. B-scan sonography along with color-coded duplex sonography is the gold standard in the diagnostic workup of these lesions, whereas histopathology is to date the gold standard for the final diagnosis. To date no single imaging technique provides the chance for an art-diagnosis with highly diagnostic accuracy. Contrast enhanced ultrasonography (CEUS) on the other hand provides information of the perfusion down to the capillary level. Currently there are only a few papers published with systematical examination of the perfusion in benign parotid gland tumors and its diagnostic significance.

PATIENTS AND METHODS

One hundred patients with a parotid gland tumor were examined. The examinations included conventional B-scan sonography, color-coded duplexsonography along with contrast enhanced ultrasonography (CEUS). B-scan sonographic parameters, i.e. echogenicity, shape, size, demarcation, and borders of a lesion along with vascularization estimated by color-coded-duplexsonography were analyzed. Analysis of quantitative CEUS parameters was performed using 8 regions of interest (ROI), which were standardized located throughout the entire tumors. The perfusion parameters were analyzed for particular tumor entities. Qualitative CEUS analysis with estimating the perfusion pattern was additionally performed.

RESULTS

Histological examination revealed benign tumors in 92 cases, with pleomorphic adenomas and Warthin´s tumors were the most frequent entities. Malignant conditions were found in 8 cases. CEUS revealed a centripetal perfusion pattern in malignant tumors significantly more frequently than in benign tumors. CEUS showed a significant heterogenic perfusion in all tumors, with a higher perfusion in the medial parts of the tumors and in some cases also in the center. Perfusion patterns of PA and WT were different. WT displayed centrifugal, centripetal, and central diffuse perfusion more often than PA, whereas in PA perfusion often was limited to the capsule or periphery. Oncocytoma had the highest perfusion values. Intraglandular cysts showed no intralesional perfusion.

CONCLUSIONS

CEUS analysis in different parts of benign tumors revealed a significant heterogeneity in tumor perfusion. Some perfusion pattern could be identified which might be characteristic for particular lesions. Based on this, the diagnostic accuracy of CEUS in the differential diagnosis of parotid gland tumors can be increased. In particular, the perfusion analysis within the tumors using ROIs located standardized throughout the entire tumor provides additional information which are important for the art diagnosis and in differentiation of tumor entity.

CT perfusion in hyper-acute ischemic stroke: the acid test for COVID-19 fear

PURPOSE

The fear of COVID-19 infection may discourage patients from going to the hospital even in case of sudden onset of disabling symptoms. There is growing evidence of the reduction of stroke admissions and higher prevalence of severe clinical presentation. Yet, no studies have investigated the perfusion pattern of acute strokes admitted during the lockdown. We aimed to evaluate the effects of the COVID-19 pandemic on hyper-acute stroke CT perfusion (CTP) pattern during the first months of the pandemic in Italy.

METHODS

In this retrospective observational study, we analyzed CTP images and clinical data of ischemic stroke patients admitted between 9 March and 2 June 2020 that underwent CTP (n = 30), to compare ischemic volumes and clinical features with stroke patients admitted during the same period in 2019 (n = 51). In particular, CTP images were processed to calculate total hypoperfused volumes, core volumes, and mismatch. The final infarct volumes were calculated on follow-up CT.

RESULTS

Significantly higher total CTP hypoperfused volume (83.3 vs 18.5 ml, p = 0.003), core volume (27.8 vs 1.0 ml, p < 0.001), and unfavorable mismatch (0.51 vs 0.91, p < 0.001) were found during the COVID-19 period compared to no-COVID-19 one. The more unfavorable perfusion pattern at admission resulted in higher infarct volume on follow-up CT during COVID-19 (35.5 vs 3.0 ml, p < 0.001). During lockdown, a reduction of stroke admissions (- 37%) and a higher prevalence of severe clinical presentation (NIHSS ≥ 10; 53% vs 36%, p = 0.029) were observed.

CONCLUSION

The results of CTP analysis provided a better insight in the higher prevalence of major severity stroke patients during the COVID-19 period.

Power-Doppler perfusion phenotype in RA patients is dependent on anti-citrullinated peptide antibody status, not on rheumatoid factor

It is not known whether there are any consistent non-serological differences between seropositive and seronegative rheumatoid arthritis, and if any, whether they depend upon rheumatoid factor (RF), anti-citrullinated peptide antibodies (ACPA), or both. In a pilot study, we showed that the two forms could be differentiated using power-Doppler sonography (PDS), and that the difference is ACPA dependent. This extended study explored whether the previous findings could be confirmed. 103 patients 51 ACPA positive (ACPA +), 52 ACPA negative (ACPA -) with active wrist arthritis were examined using PDS. By means of a temporal image series, pulsatility was evaluated over a 3-5-s period, maximum and minimum perfusion signal were determined using a computer program counting the number of coloured pixels for each frame. Maxima (P_max) and minima (P_min) were determined, and the standardized peak-to-peak amplitude sA was calculated (sA = (P_max - P_min)/P_max). This parameter was then compared for ACPA + and ACPA- patients. In addition, a multivariate regression was performed, to determine which factors influence sA. sA differed significantly between ACPA + and ACPA- patients [20% (13-26) vs. 41% (32-57), p < 0. 0001]. In the multivariate analysis, age (t = 2.5, p = 0.02) and ACPA status (t = - 4.8, p < 0.0001) were independent predictors of sA. PDS perfusion patterns are different in seropositive and seronegative RA. The difference appears to be ACPA, not RF dependent. This suggests that the underlying pathophysiological process is different in ACPA-positive and ACPA-negative RA.

Perfusion patterns of ischemic stroke on computed tomography perfusion

CT perfusion (CTP) has been applied increasingly in research of ischemic stroke. However, in clinical practice, it is still a relatively new technology. For neurologists and radiologists, the challenge is to interpret CTP results properly in the context of the clinical presentation. In this article, we will illustrate common CTP patterns in acute ischemic stroke using a case-based approach. The aim is to get clinicians more familiar with the information provided by CTP with a view towards inspiring them to incorporate CTP in their routine imaging workup of acute stroke patients.