Imaging of primary malignant bone tumors (nonhematological)

Diagnostic performance and inter-reader reliability of bone reporting and data system (Bone-RADS) on computed tomography

OBJECTIVE

To evaluate the diagnostic performance and inter-reader reliability of the Bone Reporting and Data System (Bone-RADS) for solitary bone lesions on CT.

MATERIALS AND METHODS

This retrospective analysis included 179 patients (mean age, 56 ± 18 years; 94 men) who underwent bone biopsies between March 2005 and September 2021. Patients with solitary bone lesions on CT and sufficient histopathology results were included. Two radiologists categorized the bone lesions using the Bone-RADS (1, benign; 4, malignant). The diagnostic performance of the Bone-RADS was calculated using histopathology results as a standard reference. Inter-reader reliability was calculated.

RESULTS

Bone lesions were categorized into two groups: 103 lucent (pathology: 34 benign, 12 intermediate, 54 malignant, and 3 osteomyelitis) and 76 sclerotic/mixed (pathology: 46 benign, 2 intermediate, 26 malignant, and 2 osteomyelitis) lesions. The Bone-RADS for lucent lesions had sensitivities of 95% and 82%, specificities of 11% and 11%, and accuracies of 57% and 50% for readers 1 and 2, respectively. The Bone-RADS for sclerotic/mixed lesions had sensitivities of 75% and 68%, specificities of 27% and 27%, and accuracies of 45% and 42% for readers 1 and 2, respectively. Inter-reader reliability was moderate to very good (κ = 0.744, overall; 0.565, lucent lesions; and 0.851, sclerotic/mixed lesions).

CONCLUSION

Bone-RADS has a high sensitivity for evaluating malignancy in lucent bone lesions and good inter-reader reliability. However, it has poor specificity and accuracy for both lucent and sclerotic/mixed lesions. A possible explanation is that proposed algorithms heavily depend on clinical features such as pain and history of malignancy.

Application of Machine Learning for Differentiating Bone Malignancy on Imaging: A Systematic Review

An accurate diagnosis of bone tumours on imaging is crucial for appropriate and successful treatment. The advent of Artificial intelligence (AI) and machine learning methods to characterize and assess bone tumours on various imaging modalities may assist in the diagnostic workflow. The purpose of this review article is to summarise the most recent evidence for AI techniques using imaging for differentiating benign from malignant lesions, the characterization of various malignant bone lesions, and their potential clinical application. A systematic search through electronic databases (PubMed, MEDLINE, Web of Science, and clinicaltrials.gov) was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A total of 34 articles were retrieved from the databases and the key findings were compiled and summarised. A total of 34 articles reported the use of AI techniques to distinguish between benign vs. malignant bone lesions, of which 12 (35.3%) focused on radiographs, 12 (35.3%) on MRI, 5 (14.7%) on CT and 5 (14.7%) on PET/CT. The overall reported accuracy, sensitivity, and specificity of AI in distinguishing between benign vs. malignant bone lesions ranges from 0.44–0.99, 0.63–1.00, and 0.73–0.96, respectively, with AUCs of 0.73–0.96. In conclusion, the use of AI to discriminate bone lesions on imaging has achieved a relatively good performance in various imaging modalities, with high sensitivity, specificity, and accuracy for distinguishing between benign vs. malignant lesions in several cohort studies. However, further research is necessary to test the clinical performance of these algorithms before they can be facilitated and integrated into routine clinical practice.

Deep Learning Assisted Diagnosis of Musculoskeletal Tumors Based on Contrast-Enhanced Magnetic Resonance Imaging

BACKGROUND

Misdiagnosis of malignant musculoskeletal tumors may lead to the delay of intervention, resulting in amputation or death.

PURPOSE

To improve the diagnostic efficacy of musculoskeletal tumors by developing deep learning (DL) models based on contrast-enhanced magnetic resonance imaging and to quantify the improvement in diagnostic performance obtained by using these models.

STUDY TYPE

Retrospective.

POPULATION

Three hundreds and four musculoskeletal tumors, including 212 malignant and 92 benign lesions, were randomized into the training (n = 180), validation (n = 62) and testing cohort (n = 62).

FIELD STRENGTH/SEQUENCE

A 3 T/T₁ -weighted (T₁ -w), T₂ -weighted (T₂ -w), diffusion-weighted imaging (DWI), and contrast-enhanced T1-weighted (CET₁ -w) images.

ASSESSMENT

Three DL models based, respectively, on the sagittal, coronal, and axial MR images were constructed to predict the malignancy of tumors. Blinded to the prediction results, a group of specialists made independent initial diagnoses for each patient by reading all image sequences. One month after the initial diagnoses, the same group of doctors made another round of diagnoses knowing the malignancy of each tumor predicted by the three models. The reference standard was the pathological diagnosis of malignancy.

STATISTICAL TESTS

Sensitivity, specificity, and accuracy (all with 95% confidential intervals [CI]) corresponding to each diagnostic test were computed. Chi-square tests were used to assess the differences in those parameters with and without DL models. A P value < 0.05 was considered statistically significant.

RESULTS

The developed models significantly improved the diagnostic sensitivities of two oncologists by 0.15 (95% CI: 0.06-0.24) and 0.36 (95% CI: 0.24-0.28), one radiologist by 0.12 (95% CI: 0.04-0.20), and three of the four orthopedists, respectively, by 0.12 (95% CI: 0.04-0.20), 0.29 (95% CI: 0.18-0.40), and 0.23 (95% CI: 0.13-0.33), without impairing any of their diagnostic specificities (all P > 0.128).

DATA CONCLUSION

The DL models developed can significantly improve the performance of doctors with different training and experience in diagnosing musculoskeletal tumors.

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 2.

Imaging of Pediatric Bone Tumors and Tumor-like Lesions

Bone lesions are commonly seen when reporting pediatric skeletal imaging. Distinguishing aggressive from nonaggressive lesions is essential in making the diagnosis. Not all aggressive lesions are neoplastic; indeed, osteomyelitis frequently presents with aggressive appearances and is far more commonly seen in the pediatric population than neoplastic lesions. In this article, we discuss an approach for the diagnosis of pediatric bone tumors and tumor-like conditions. The most common pediatric benign and malignant bone tumors are discussed in more detail.

Solitary bone tumor imaging reporting and data system (BTI-RADS): initial assessment of a systematic imaging evaluation and comprehensive reporting method

OBJECTIVES

Identify the most pertinent imaging features for solitary bone tumor characterization using a multimodality approach and propose a systematic evaluation system.

METHODS

Data from a prospective trial, including 230 participants with histologically confirmed bone tumors, typical "do not touch" lesions, and stable chondral lesions, were retrospectively evaluated. Clinical data, CT, and MR imaging features were analyzed by a musculoskeletal radiologist blinded to the diagnosis using a structured report. The benign-malignant distribution of lesions bearing each image feature evaluated was compared to the benign-malignant distribution in the study sample. Benign and malignant indicators were identified. Two additional readers with different expertise levels independently evaluated the study sample.

RESULTS

The sample included 140 men and 90 women (mean age 40.7 ± 18.3 years). The global benign-malignant distribution was 67-33%. Seven imaging features reached the criteria for benign indicators with a mean frequency of benignancy of 94%. Six minor malignant indicators were identified with a mean frequency of malignancy of 60.5%. Finally, three major malignant indicators were identified (Lodwick-Madewell grade III, aggressive periosteal reaction, and suspected metastatic disease) with a mean frequency of malignancy of 82.4%. A bone tumor imaging reporting and data system (BTI-RADS) was proposed. The reproducibility of the BTI-RADS was considered fair (kappa = 0.67) with a mean frequency of malignancy in classes I, II, III, and IV of 0%, 2.2%, 20.1%, and 71%, respectively.

CONCLUSION

BTI-RADS is an evidence-based systematic approach to solitary bone tumor characterization with a fair reproducibility, allowing lesion stratification in classes of increasing malignancy frequency.

TRIAL REGISTRATION

Clinical trial number NCT02895633 .

KEY POINTS

• The most pertinent CT and MRI criteria allowing bone tumor characterization were defined and presented. • Lodwick-Madewell grade III, aggressive periosteal reaction, and suspected metastatic disease should be considered major malignant indicators associated with a frequency of malignancy over 75%. • The proposed evidence-based multimodality reporting system stratifies solitary bone tumors in classes with increasing frequencies of malignancy.

A CT-based radiomics nomogram for distinguishing between benign and malignant bone tumours

BACKGROUND

We sought to evaluate the performance of a computed tomography (CT)-based radiomics nomogram we devised in distinguishing benign from malignant bone tumours.

METHODS

Two hundred and six patients with bone tumours were spilt into two groups: a training set (n = 155) and a validation set (n = 51). A feature extraction process based on 3D Slicer software was used to extract the radiomics features from unenhanced CT images, and least absolute shrinkage and selection operator logistic regression was used to calculate the radiomic score to generate a radiomics signature. A clinical model comprised demographics and CT features. A radiomics nomogram combined with the clinical model and the radiomics signature was constructed. The performance of the three models was comprehensively evaluated from three aspects: identification ability, accuracy, and clinical value, allowing for generation of an optimal prediction model.

RESULTS

The radiomics nomogram comprised clinical and radiomics signature features. The nomogram model displayed good performance in training and validation sets with areas under the curve of 0.917 and 0.823, respectively. The areas under the curve, decision curve analysis, and net reclassification improvement showed that the radiomics nomogram model could obtain better diagnostic performance than the clinical model and achieve greater clinical net benefits than the clinical and radiomics signature models alone.

CONCLUSIONS

We constructed a combined nomogram comprising a clinical model and radiomics signature as a noninvasive preoperative prediction method to distinguish between benign and malignant bone tumours and assist treatment planning.

Shoulder Tumor/Tumor-Like Lesions: What to Look for

This article discusses the most common tumor and tumor-like lesions arising at the shoulder. Osseous tumors of the shoulder rank second in incidence to those at the knee joint and include benign osteochondromas and myeloma or primary malignant lesions, such as osteosarcoma or chondrosarcomas. Soft tissue tumors are overwhelmingly benign, with lipomas predominating, although malignant lesions, such as liposarcomas, can occur. Numerous tumor-like lesions may arise from the joints or bursae, due to either underlying arthropathy and synovitis (eg, rheumatoid arthritis and amyloid) or related to conditions, including tenosynovial giant cell tumor and synovial osteochondromatosis.

Imaging Features of Bone Tumors: Conventional Radiographs and MR Imaging Correlation

Imaging bone tumors often causes uncertainty, especially outside dedicated sarcoma treatment centers. Conventional radiography remains the backbone of bone tumor diagnostics, but MR imaging has a role. Radiographs are crucial for assessing the tumor matrix and aggressiveness. MR imaging is the best modality for local staging. This article reviews semiological aspects of bone tumors: patient age, tumor localization, pattern of bone destruction/margins, aggressiveness, growth speed, matrix formation, periosteal reaction, cortical involvement, size, and number of lesions. All aspects are discussed in terms of their appearance on radiographs and MR imaging, with a focus on the correlation between the 2 modalities.

Bone Tumor Diagnosis Using a Naïve Bayesian Model of Demographic and Radiographic Features

Because many bone tumors have a variety of appearances and are uncommon, few radiologists develop sufficient expertise to guide optimal management. Bayesian inference can guide decision-making by computing probabilities of multiple diagnoses to generate a differential. We built and validated a naïve Bayes machine (NBM) that processes 18 demographic and radiographic features. We reviewed over 1664 analog radiographic cases of bone tumors and selected 811 cases (66 diagnoses) for annotation using a quantitative imaging platform. Leave-one-out cross validation was performed. Primary accuracy was defined as the correct pathological diagnosis as the top machine prediction. Differential accuracy was defined as whether the correct pathological diagnosis was within the top three predictions. For the 29 most common diagnoses (710 cases), primary accuracy was 44%, and differential accuracy was 60%. For the top 10 most common diagnoses (478 cases), primary accuracy was 62%, and differential accuracy was 80%. The machine returned relevant diagnoses for the majority of unknown test cases and may be a feasible alternative to machine learning approaches such as deep neural networks or support vector machines that typically require larger training data (our model required a minimum of five samples per diagnosis) and are "black boxes" (our model can provide details of probability calculations to identify features that most significantly contribute to truth diagnoses). Finally, our Bayes model was designed to scale and "learn" from external data, enabling incorporation of outside knowledge such as Dahlin's Bone Tumors, a reference of anatomic and demographic statistics of more than 10,000 tumors.

[Treatment of adamantinoma of femur with limb preservation. A case report and review of the literature]

BACKGROUND

Adamantinoma is a rare lesion of low-grade malignancy, and represents 1% of malignant bone tumours of bones, and is mainly located in two regions of the body, jaw (ameloblastoma), and lower extremities. The treatment of choice is surgery due to it being a radio- and chemotherapy-resistant neoplasia.

CLINICAL CASE

A 39 year old male with a history of neonatal hydrocephalus with moderate psychomotor retardation. He began with pain in the posterior region of the left thigh for one year before admission, which was managed as posterior radicular syndrome. He had sudden intense pain on walking, that led him to fall over. In the examination, left pelvic limb with deformity in the distal third with increase in volume in the thigh, with pain to palpation, and presence of crackles in the distal third of the femur. A biopsy of the thigh was performed, with subsequent local wide excision + replacement of bone with cadaver bone and a central medullary nail. The final diagnosis was adamantinoma of femur.

CONCLUSION

The adamantinomas are rare tumours. It is important to recognise this type of tumor from the beginning, since its prognosis is excellent in initial stages. It is important to have free margins as survival is very high.

Multi-detector computed tomography in evaluating locally aggressive and malignant bone tumours

OBJECTIVE

To evaluate the ability of Multi-Detector Computed Tomography in preoperative evaluation of locally aggressive and malignant bone tumours in correlation with histopathological findings.

MATERIALS AND METHODS

Twenty patients suspected of malignant bone tumours on the basis of their clinical profile were selected. Following a plain radiograph evaluation, all of them were subjected to CT scan examination. Multi Planar Reconstruction (MPR) was done in sagittal and coronal planes and also three-dimensional Volume Rendering (VR) and Maximum Intensity Projection (MIP) images were obtained.

RESULTS

Of the 20 patients, 18 underwent surgery, and their histopathological findings were compared and correlated with MDCT findings. MDCT was 92.8% sensitive and 100% specific in determining the vascularity of the tumour and also can detect displacement/ encasement/ involvement of adjacent vessels. It has a sensitivity and specificity of 100% in determining cortical break, calcification and periosteal reaction. However, it is less sensitive in detecting joint involvement. Post contrast enhancement gives details of the extent of the soft tissue component.

CONCLUSION

Although MRI is a preferred modality in preoperative evaluation of bone tumours, CT may be used an alternative in case of non-availability of MRI, which has faster acquisition time and better resolution. Using three dimensional MPR imaging, the location and extent of the tumour can be studied. It is also useful in determining cortical discontinuity, periosteal reaction, and calcification. By virtue of MIP and VR imaging, vascularity of the tumour and its relationship with the adjacent vasculature can be established. However, it is inferior to MRI in soft tissue characterization and has poor sensitivity in detecting marrow and joint involvement.

[Imaging diagnostics of bone sarcomas]

BACKGROUND

Bone tumors and especially bone sarcomas are rare lesions of the skeletal system in comparison to the much more frequently occurring bone metastases. Despite the relative rarity they are important differential diagnoses of bone lesions.

OBJECTIVE

The aim of this article is to give the reader an insight into the fundamentals of the primary imaging of bone sarcomas and to illustrate this with the help of two examples (e.g. osteosarcoma and chondrosarcoma).

RESULTS

The foundation of the imaging of bone sarcomas is the radiograph in two planes. This method delivers important information on bone tumors. This information should be analyzed with the help of the Lodwick classification, the configuration of periosteal reactions and a possible reaction of the cortex. A possible tumor matrix and the localization within the skeleton or within long bones also provide important information for differential diagnostic delimitation. Magnetic resonance imaging (MRI) with specific adapted bone tumor sequences allows an exact local staging of a bone sarcoma. In addition to local imaging a compartmental MRI which illustrates the entire extent of tumor-bearing bone and the adjacent joints should be performed to rule out possible skip lesions. The most common distant metastases of osteosarcoma and chondrosarcoma occur in the lungs; therefore, a computed tomography (CT) of the chest is part of staging. Other imaging methods, such as CT of the tumor, positron emission tomography CT (PET-CT), bone scan and whole body MRI supplement the imaging depending on tumor type.

Diagnosing bone lesions: a practical non-pathology centric formulaic approach

Bone lesions are perceived to be some of the most difficult lesions that pathologists encounter. The reasons for this are multiple and include lack of experience/familiarity, the need to rely heavily on non-pathology information and data, and the fact that many lesions are associated with either procedures or treatments with significant morbidity. However, in fact, the majority of bone lesions can be accurately assessed on the basis of data not directly related to traditional pathologic based assessment. In order to achieve this state, the pathologist must understand the consistent clinical parameters of most bone lesions, including their clinical presentation, the bone involved, particularly the anatomic site of the bone involved, and a fundamental, basic understanding of imaging studies, especially the plain radiograph. Once these principles are understood and mastered, the pathologist can easily diagnose most bone lesions, using traditional pathologic assessment to confirm the diagnosis.

Imaging evaluation of musculoskeletal tumors

In this chapter, we review different imaging modalities, including radiography, computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and nuclear medicine scintigraphy, and their application to musculoskeletal neoplasm. Advantages and limitations of each modality are reviewed, and suggestions for imaging approach are provided.

Polymorphisms in the MDM2 gene and risk of malignant bone tumors: a meta-analysis

There are several studies published to assess the associations of murine double minute 2 (MDM2) genetic polymorphisms with risk of malignant bone tumors, but they reported contradictory results and failed to confirm a strong and consistent association. To assess the evidence regarding the associations of MDM2 genetic polymorphisms with the risk of malignant bone tumors, we conducted a meta-analysis of epidemiological studies. The pooled odds ratio (OR) with its 95% confidence intervals (95% CI) was used to assess these possible associations. Four studies with a total of 3,958 individuals were finally included the meta-analysis. Meta-analysis of two studies on MDM2 SNP309 polymorphism showed that MDM2 SNP309 polymorphism was associated with an increased risk of malignant bone tumors (G versus T: OR = 1.72, 95% CI 1.35-2.20, P < 0.001; GG versus TT: OR = 2.64, 95% CI 1.59-4.39, P < 0.001; GG/GT versus TT: OR = 1.87, 95% CI 1.33-2.62, P < 0.001; GG versus TT/GT: OR = 2.20, 95% CI 1.38-3.51, P = 0.001). Meta-analysis of those two studies on MDM2 rs1690916 polymorphism showed that MDM2 rs1690916 minor allele A was associated with decreased risk of malignant bone tumors (OR = 0.60, 95% CI 0.46-0.77, P < 0.001). Meta-analyses of available data show that there are significant associations of MDM2 SNP309 polymorphism and MDM2 rs1690916 polymorphism with malignant bone tumors.