Comparison of Commercial AI Software Performance for Radiograph Lung Nodule Detection and Bone Age Prediction.

Background Multiple commercial artificial intelligence (AI) products exist for assessing radiographs; however, comparable performance data for these algorithms are limited. Purpose To perform an independent, stand-alone validation of commercially available AI products for bone age prediction based on hand radiographs and lung nodule detection on chest radiographs. Materials and Methods This retrospective study was carried out as part of Project AIR. Nine of 17 eligible AI products were validated on data from seven Dutch hospitals. For bone age prediction, the root mean square error (RMSE) and Pearson correlation coefficient were computed. The reference standard was set by three to five expert readers. For lung nodule detection, the area under the receiver operating characteristic curve (AUC) was computed. The reference standard was set by a chest radiologist based on CT. Randomized subsets of hand (n = 95) and chest (n = 140) radiographs were read by 14 and 17 human readers, respectively, with varying experience. Results Two bone age prediction algorithms were tested on hand radiographs (from January 2017 to January 2022) in 326 patients (mean age, 10 years ± 4 [SD]; 173 female patients) and correlated strongly with the reference standard (r = 0.99; P < .001 for both). No difference in RMSE was observed between algorithms (0.63 years [95% CI: 0.58, 0.69] and 0.57 years [95% CI: 0.52, 0.61]) and readers (0.68 years [95% CI: 0.64, 0.73]). Seven lung nodule detection algorithms were validated on chest radiographs (from January 2012 to May 2022) in 386 patients (mean age, 64 years ± 11; 223 male patients). Compared with readers (mean AUC, 0.81 [95% CI: 0.77, 0.85]), four algorithms performed better (AUC range, 0.86-0.93; P value range, <.001 to .04). Conclusions Compared with human readers, four AI algorithms for detecting lung nodules on chest radiographs showed improved performance, whereas the remaining algorithms tested showed no evidence of a difference in performance. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Omoumi and Richiardi in this issue.

Normative mammillary body volumes: From the neonatal period to young adult.

The mammillary bodies may be small, but they have an important role in encoding complex memories. Mammillary body pathology often occurs following thiamine deficiency but there is increasing evidence that the mammillary bodies are also compromised in other neurological conditions and in younger ages groups. For example, the mammillary bodies are frequently affected in neonates with hypoxic-ischemic encephalopathy. At present, there is no normative data for the mammillary bodies in younger groups making it difficult to identify abnormalities in neurological disorders. To address this, the present study set out to develop a normative dataset for neonates and for children to young adult. A further aim was to determine whether there were laterality or sex differences in mammillary body volumes. Mammillary body volumes were obtained from MRI scans from 506 participants across two datasets. Measures for neonates were acquired from the Developing Human Connectome Project database (156 male; 100 female); volumes for individuals aged 6-24 were acquired from the NICHE database (166 males; 84 females). Volume measurements were acquired using a semi-automated multi-atlas segmentation approach. Mammillary body volumes increased up to approximately 15 years-of-age. The left mammillary body was marginally, but significantly, larger than the right in the neonates with a similar pattern in older children/young adults. In neonates, the mammillary bodies in males were slightly bigger than females but no sex differences were present in older children/young adults. Given the increasing presentation of mammillary body pathology in neonates and children, these normative data will enable better assessment of the mammillary bodies in healthy and at-risk populations.

Mammillary body injury in neonatal encephalopathy: a multicentre, retrospective study.

BACKGROUND: The mammillary bodies (MBs) have repeatedly been shown to be critical for memory, yet little is known about their involvement in numerous neurological conditions linked to memory impairments, including neonatal encephalopathy. METHODS: We implemented a multicentre retrospective study, assessing magnetic resonance scans of 219 infants with neonatal encephalopathy who had undergone hypothermia treatment in neonatal intensive care units located in the Netherlands and Italy. RESULTS: Abnormal MB signal was observed in ~40% of infants scanned; in half of these cases, the brain appeared otherwise normal. MB involvement was not related to the severity of encephalopathy or the pattern/severity of hypoxic-ischaemic brain injury. Follow-up scans were available for 18 cases with abnormal MB signal; in eight of these cases, the MBs appeared severely atrophic. CONCLUSIONS: This study highlights the importance of assessing the status of the MBs in neonatal encephalopathy; this may require changes to scanning protocols to ensure that the slices are sufficiently thin to capture the MBs. Furthermore, long-term follow-up of infants with abnormal MB signal is needed to determine the effects on cognition, which may enable the use of early intervention strategies. Further research is needed to assess the role of therapeutic hypothermia in MB involvement in neonatal encephalopathy. IMPACT: The MBs are particularly sensitive to hypoxia in neonates. Current hypothermia treatment provides incomplete protection against MB injury. MB involvement is likely overlooked as it can often occur when the rest of the brain appears normal. Given the importance of the MBs for memory, it is necessary that this region is properly assessed in neonatal encephalopathy. This may require improvements in scanning protocols.

Recurrence of Ewing sarcomas of the chest wall.

BACKGROUND: Ewing sarcomas (ES) of the chest wall are rare. Local recurrences occur in approximately 20% of these patients; however literature on this topic is scarce. Our aim was to analyze the influence of the extent of surgical resection on outcome, and to find positive prognostic factors for survival. PROCEDURE: A retrospective analysis of all patients who were under 18 and treated between 1977 and 2006 for Ewing sarcoma of the chest wall was performed. RESULTS: Twenty-seven patients with ES were diagnosed at our institution; 14 males and 13 females. Twenty-three patients underwent local resection, which was complete in 13. Nine patients had a wide resection, including more ribs than just the affected one. Complete remission (CR) was achieved in 14/27 patients. Recurrence was observed in 13 patients: local recurrence (LR) and distant metastasis (DM) both in 7. One of the patients had both LR and DM. All patients with LR and two patients with DM only died of disease. Overall survival after 10 and 20 years was 73% and 47%, respectively. Complete resection, wide resection, gender, or age were no significant factors affecting outcome. CONCLUSIONS: Relapse carries a high mortality risk and none of the patients with LR eventually survived. Wide resection of at least a complete normal adjacent rib above and below the thoracic wall tumor as a fixed rule seems unnecessary.