Encoder-decoder models for chest X-ray report generation perform no better than unconditioned baselines

High quality radiology reporting of chest X-ray images is of core importance for high-quality patient diagnosis and care. Automatically generated reports can assist radiologists by reducing their workload and even may prevent errors. Machine Learning (ML) models for this task take an X-ray image as input and output a sequence of words. In this work, we show that ML models for this task based on the popular encoder-decoder approach, like ‘Show, Attend and Tell’ (SA&T) have similar or worse performance than models that do not use the input image, called unconditioned baseline. An unconditioned model achieved diagnostic accuracy of 0.91 on the IU chest X-ray dataset, and significantly outperformed SA&T (0.877) and other popular ML models (p-value < 0.001). This unconditioned model also outperformed SA&T and similar ML methods on the BLEU-4 and METEOR metrics. Also, an unconditioned version of SA&T obtained by permuting the reports generated from images of the test set, achieved diagnostic accuracy of 0.862, comparable to that of SA&T (p-value ≥ 0.05).

Structured reporting in radiology: a systematic review to explore its potential

Objectives Structured reporting (SR) in radiology reporting is suggested to be a promising tool in clinical practice. In order to implement such an emerging innovation, it is necessary to verify that radiology reporting can benefit from SR. Therefore, the purpose of this systematic review is to explore the level of evidence of structured reporting in radiology. Additionally, this review provides an overview on the current status of SR in radiology. Methods A narrative systematic review was conducted, searching PubMed, Embase, and the Cochrane Library using the syntax ‘radiol*’ AND ‘structur*’ AND ‘report*’. Structured reporting was divided in SR level 1, structured layout (use of templates and checklists), and SR level 2, structured content (a drop-down menu, point-and-click or clickable decision trees). Two reviewers screened the search results and included all quantitative experimental studies that discussed SR in radiology. A thematic analysis was performed to appraise the evidence level. Results The search resulted in 63 relevant full text articles out of a total of 8561 articles. Thematic analysis resulted in 44 SR level 1 and 19 level 2 reports. Only one paper was scored as highest level of evidence, which concerned a double cohort study with randomized trial design. Conclusion The level of evidence for implementing SR in radiology is still low and outcomes should be interpreted with caution. Key Points • Structured reporting is increasingly being used in radiology, especially in abdominal and neuroradiological CT and MRI reports. • SR can be subdivided into structured layout (SR level 1) and structured content (SR level 2), in which the first is defined as being a template in which the reporter has to report; the latter is an IT-based manner in which the content of the radiology report can be inserted and displayed into the report. • Despite the extensive amount of research on the subject of structured reporting, the level of evidence is low.

Hip surgery and radiology reporting for children with cerebral palsy prior to initiation of a hip surveillance program

PURPOSE: Hip displacement impacts quality of life for many children with cerebral palsy (CP). While early detection can help avoid dislocation and late-stage surgery, formalized surveillance programs are not ubiquitous. This study aimed to examine: 1) surgical practices around pediatric hip displacement for children with CP in a region without formalized hip surveillance; and 2) utility of MP compared to traditional radiology reporting for quantifying displacement. METHODS: A retrospective chart review examined hip displacement surgeries performed on children with CP between 2007–2016. Surgeries were classified as preventative, reconstructive, or salvage. Pre- and post-operative migration percentage (MP) was calculated for available radiographs using a mobile application and compared using Wilcoxon Signed Ranks test. MPs were also compared with descriptions in the corresponding radiology reports using directed and conventional content analyses. RESULTS: Data from 67 children (115 surgical hips) was included. Primary surgery types included preventative (63.5% hips), reconstructive (36.5%), or salvage (0%). For the 92 hips with both radiology reports and radiographs available, reports contained a range of descriptors that inconsistently reflected the retrospectively-calculated MPs. CONCLUSION: Current radiology reporting practices do not appear to effectively describe hip displacement for children with CP. Therefore, standardized reporting of MP is recommended.

Reversing the Friday peak in metastatic cord compression referrals: Not as simple as previously thought?

e14050 Background: Historically, metastatic spinal cord compression (MSCC) referrals trend towards a Friday peak in incidence (Koiter E, Radioth Onc 2013). However, data from a single, tertiary centre in the UK showed a reversal in the Friday peak (Adeleke S, Annals of Oncology 2020). This was attributed to early case referrals and quicker treatment decisions. In this new study, we explored whether a similar pattern was apparent in multiple district general hospital (DGH) settings and attempt to identify underlying causes. DGHs manage a larger proportion of cancer patients in the UK. Methods: 1,069 patients between 1 Jan 2015 and 31 Dec 2020 were identified across 4 hospitals in Kent, UK with a population of 1.6 million people. 220, 181, 182, 159, 134 and 193 MSCC patients were identified annually (2015-2020). Commonest cancers were prostate (24.1%), lung (19.3%) and breast (12.3%). Thoracic and lumbar regions constituted 80% of MSCC sites. Kruskal Wallis was used to compare differences in referrals across weekdays. Data was then dichotomised to Fridays only vs. other days of the week combined, as previously reported (De Bono B, Acta Neurochir 2019). Chi squared was used to compare frequency of referrals between the two groups. Chi squared goodness of fit test was conducted to detect if Friday reflected the day with highest referrals across the week. Results: Across the region, 2015 saw the highest number of Friday referrals relative to other days, p= 0.002. Friday referrals continued to drop, year on year, until 2018 with a corresponding increase in mid-week referrals. After 2018, there was a return in trend to a further Friday peak across the region, though p= 0.836. On an individual hospital basis, the persistent Friday peak in the region was driven by two hospitals. Having a 7-day acute oncology service (AOS), 7-day radiology reporting and single referral point of contact in the department, were factors identified that kept the referrals across the week uniform. On another note, a substantial shift towards a single 8Gy fraction vs. 20Gy in 5 fractions was observed across the region. This change coincided with SCORAD III data (Hoskin P, ASCO 2017) and demonstrates adherence to evidence-based practice in the region. Conclusions: This large multi-centre retrospective study shows a differential referral pattern in the region, with hospitals with 7-day AOS/Radiology reporting and single point of referral (e.g, similar to MSCC coordinator role) having a quicker treatment turnaround and uniform referrals across the week. The MSCC coordinator has been shown to streamline service, ensure timely decision-making and improved survival outcomes (Richards L, Spine J 2017). The role is recommended by NICE UK. DGHs should consider appointing an MSCC coordinator when designing/auditing their service. The shift towards single 8Gy fraction can provide a ‘one-stop’ service where patients are scanned, planned and treated on the same day.

Computer-Assisted Reporting and Decision Support in Standardized Radiology Reporting for Cancer Imaging

PURPOSE Recent advances in structured reporting are providing an opportunity to enhance cancer imaging assessment to drive value-based care and improve patient safety. METHODS The computer-assisted reporting and decision support (CAR/DS) framework has been developed to enable systematic ingestion of guidelines as clinical decision structured reporting tools embedded within the radiologist's workflow. RESULTS CAR/DS tools can reduce the radiology reporting variability and increase compliance with clinical guidelines. The lung cancer use-case is used to describe various scenarios of a cancer imaging structured reporting pathway, including incidental findings, screening, staging, and restaging or continued care. Various aspects of these tools are also described using cancer-related examples for different imaging modalities and applications such as calculators. Such systems can leverage artificial intelligence (AI) algorithms to assist with the generation of structured reports and there are opportunities for new AI applications to be created using the structured data associated with CAR/DS tools. CONCLUSION These AI-enabled systems are starting to allow information from multiple sources to be integrated and inserted into structured reports to drive improvements in clinical decision support and patient care.