An imaging protocol for the management of central cartilage tumours of the proximal fibula

An evaluation of AI generated literature reviews in musculoskeletal radiology

PURPOSE

The use of artificial intelligence (AI) tools to aid in summarizing information in medicine and research has recently garnered a huge amount of interest. While tools such as ChatGPT produce convincing and naturally sounding output, the answers are sometimes incorrect. Some of these drawbacks, it is hoped, can be avoided by using programmes trained for a more specific scope. In this study we compared the performance of a new AI tool (the-literature.com) to the latest version OpenAI's ChatGPT (GPT-4) in summarizing topics that the authors have significantly contributed to.

METHODS

The AI tools were asked to produce a literature review on 7 topics. These were selected based on the research topics that the authors were intimately familiar with and have contributed to through their own publications. The output produced by the AI tools were graded on a 1-5 Likert scale for accuracy, comprehensiveness, and relevance by two fellowship trained consultant radiologists.

RESULTS

The-literature.com produced 3 excellent summaries, 3 very poor summaries not relevant to the prompt, and one summary, which was relevant but did not include all relevant papers. All of the summaries produced by GPT-4 were relevant, but fewer relevant papers were identified. The average Likert rating was for the-literature was 2.88 and 3.86 for GPT-4. There was good agreement between the ratings of both radiologists (ICC = 0.883).

CONCLUSION

Summaries produced by AI in its current state require careful human validation. GPT-4 on average provides higher quality summaries. Neither tool can reliably identify all relevant publications.

Intraosseous conventional central chondrosarcoma does not metastasise irrespective of grade in pelvis, scapula and in long bone locations

Background

Histological grade has been regarded as the most important prognostic factor in conventional central chondrosarcoma. To evaluate whether the presence of an extraosseous tumour component is associated with a decreased metastasis-free survival or disease-specific survival and alternatively to develop a simple prognostic and clinical decision-making tool.

Material and methods

We searched two prospectively maintained international sarcoma centre databases for primary non metastatic central conventional chondrosarcomas of all grades in pelvis, scapula or long bone location, undergoing curative treatment, diagnosed between 2000 and 2020. Pre-treatment MRI scans were reviewed for the presence of an extraosseous mass. The metastasis-free survival (MFS) and disease-specific survival (DSS) were estimated by the Kaplan-Meier method from surgery to event, death or last follow-up.

Results

336 patients were identified between 2000 and 2020, undergoing surgical treatment for conventional central chondrosarcoma. 111 patients (33 %) had grade 1 tumours, 149 patients (44 %) had grade 2, and 76 patient (23 %) had grade 3 chondrosarcomas determined as the highest grade in the final resected specimen. An extraosseous soft tissue component was more frequent in higher grade tumours (p < 0.001) and present in 200 cases (60 %). None of the patients with an intraosseous tumour developed metastases or died of the disease. For patients with extraosseous tumour component, MFS was 92 % (95 % CI, 96-100) at 2-years and 74 % (95 % CI, 67-81) at 10-years and DSS was 91 % (95 % CI, 87-95) at 2-years and 75 % (95 % CI, 68-82) at 10-years. The MFS and DSS was significantly different (p < 0.001) for those patients with or without an extraosseous tumour component, irrespective of grade or anatomical location.

Discussion

The results of this study has shown that the metastatic potential of intraosseous conventional central chondrosarcoma is negligible. The presence of an extraosseous soft tissue component may be used for prognostication and to guide treatment pathways for patients with central cartilage tumours.

Benign incidental do-not-touch bone lesions

Benign bone lesions may occasionally be incidentally detected on radiographs and are also increasingly found on CT or MRI performed for other clinical indications. Although mostly asymptomatic or associated with minor symptoms, these lesions may simulate true pathological lesions, causing problems in diagnosis. For instance, asymptomatic benign bone lesions can be misinterpreted as metastasis when incidentally encountered in a patient with known cancer. Recognising these entities as "do-not-touch" lesions helps avoid unnecessary further investigation or harmful intervention. In this review, we highlight three groups of bone incidentalomas found in adults, namely: osteolytic lesions, osteoblastic lesions, and bone protuberances. We aim to review the key imaging features of selected common and less common conditions in these three groups, so as to help radiologists confidently identify these benign do-not-touch lesions and to distinguish them from more sinister pathological lesions.

Prevalence of Enchondromas of the Proximal Femur in Adults as an Incidental Finding on MRI of the Pelvis

Objective  To determine the prevalence of enchondroma in adults as an incidental finding in the proximal femur on magnetic resonance imaging (MRI). Materials and Methods  A retrospective review of the MRI scans of the pelvis in a series of adult patients was conducted. All presented with nononcological musculoskeletal complaints. The site, size, and appearances of the enchondromas were identified according to criteria from previous studies. Results  A total of 1,209 proximal femora in 610 patients were reviewed and a total of 9 enchondromas were identified. These ranged from 0.6 to 2.5 cm in length (mean 1.3 cm). None showed aggressive features suggestive of malignancy. Three cases (33%) underwent follow-up MRI scans which showed no change in size or morphology. Conclusion  The prevalence on MRI of incidental enchondromas arising in the proximal femur is 0.7%. This is three to four times less common than seen in the proximal humerus and around the knee.

The changing face of central chondrosarcoma of bone. One UK-based orthopaedic oncology unit's experience of 33 years referrals

AIM

To ascertain the changing incidence over time of the three commonest primary sarcomas of bone. Data obtained with particular reference to central chondrosarcoma from the annual referral rate to a large UK-based specialist orthopaedic oncology unit. To discuss how the "barnyard pen" analogy of cancers previously applied to certain commoner cancers can also be applicable to central chondrosarcoma (CS) of bone.

MATERIALS AND METHODS

A retrospective review was conducted of a computerised database identifying all central cartilage tumours (CCT) of bone, including enchondroma and CS subtypes, between 1985 and 2018. These were compared with the referrals of the other two commonest primary sarcomas of bone, osteosarcoma and Ewing sarcoma.

RESULTS

There was a total of 1507 CS showing a 68% overall increase in annual referral rate/incidence over the study period. 68% cases were the borderline malignant lesions now known as atypical cartilaginous tumour (ACT). The annual referral rate/incidence of this entity increased by 194% over the 30 years. Whereas, the annual referral rate/incidence for osteosarcoma and Ewing sarcoma was static for the past 20 years.

CONCLUSION

The annual incidence of central CS of bone showed a marked increase over the 33-year period as compared with both osteosarcoma and Ewing sarcoma. This is especially in the ACT category and is thought to be due to the increased provision of MRI scanning flagging up a rise in incidental findings. The spectrum of CCTs from benign to highly malignant elegantly fits the "barn yard" pen analogy and could prove useful as an explanatory tool for patients and clinicians unfamiliar with these diseases.