Automatic detection, segmentation, and classification of primary bone tumors and bone infections using an ensemble multi-task deep learning framework on multi-parametric MRIs: a multi-center study.

OBJECTIVES: To develop an ensemble multi-task deep learning (DL) framework for automatic and simultaneous detection, segmentation, and classification of primary bone tumors (PBTs) and bone infections based on multi-parametric MRI from multi-center. METHODS: This retrospective study divided 749 patients with PBTs or bone infections from two hospitals into a training set (N = 557), an internal validation set (N = 139), and an external validation set (N = 53). The ensemble framework was constructed using T1-weighted image (T1WI), T2-weighted image (T2WI), and clinical characteristics for binary (PBTs/bone infections) and three-category (benign/intermediate/malignant PBTs) classification. The detection and segmentation performances were evaluated using Intersection over Union (IoU) and Dice score. The classification performance was evaluated using the receiver operating characteristic (ROC) curve and compared with radiologist interpretations. RESULT: On the external validation set, the single T1WI-based and T2WI-based multi-task models obtained IoUs of 0.71 ± 0.25/0.65 ± 0.30 for detection and Dice scores of 0.75 ± 0.26/0.70 ± 0.33 for segmentation. The framework achieved AUCs of 0.959 (95%CI, 0.955-1.000)/0.900 (95%CI, 0.773-0.100) and accuracies of 90.6% (95%CI, 79.7-95.9%)/78.3% (95%CI, 58.1-90.3%) for the binary/three-category classification. Meanwhile, for the three-category classification, the performance of the framework was superior to that of three junior radiologists (accuracy: 65.2%, 69.6%, and 69.6%, respectively) and comparable to that of two senior radiologists (accuracy: 78.3% and 78.3%). CONCLUSION: The MRI-based ensemble multi-task framework shows promising performance in automatically and simultaneously detecting, segmenting, and classifying PBTs and bone infections, which was preferable to junior radiologists. CLINICAL RELEVANCE STATEMENT: Compared with junior radiologists, the ensemble multi-task deep learning framework effectively improves differential diagnosis for patients with primary bone tumors or bone infections. This finding may help physicians make treatment decisions and enable timely treatment of patients. KEY POINTS: • The ensemble framework fusing multi-parametric MRI and clinical characteristics effectively improves the classification ability of single-modality models. • The ensemble multi-task deep learning framework performed well in detecting, segmenting, and classifying primary bone tumors and bone infections. • The ensemble framework achieves an optimal classification performance superior to junior radiologists' interpretations, assisting the clinical differential diagnosis of primary bone tumors and bone infections.

Musculoskeletal infections through direct inoculation.

Musculoskeletal infections consist of different clinical conditions that are commonly encountered in daily clinical settings. As clinical findings and even laboratory tests cannot always be specific, imaging plays a crucial role in the diagnosis and treatment of these cases. Musculoskeletal infections most commonly occur secondary to direct inoculation into the skin involuntarily affected by trauma, microorganism, foreign bodies, or in diabetic ulcers; direct infections can also occur from voluntary causes due to surgery, vaccinations, or other iatrogenic procedures. Hematogenous spread of infection from a remote focus can also be a cause for musculoskeletal infections. Risk factors for soft tissue and bone infections include immunosuppression, old age, corticosteroid use, systemic illnesses, malnutrition, obesity, and burns. Most literature discusses musculoskeletal infections according to the diagnostic tools or forms of infection seen in different soft tissue anatomical planes or bones. This review article aims to evaluate musculoskeletal infections that occur due to direct inoculation to the musculoskeletal tissues, by focusing on the traumatic mechanism with emphasis on the radiological findings.

Diagnosis of peripheral bone and prosthetic joint infections: overview on the consensus documents by the EANM, EBJIS, and ESR (with ESCMID endorsement).

OBJECTIVES: Peripheral bone infection (PBI) and prosthetic joint infection (PJI) are two different infectious conditions of the musculoskeletal system. They have in common to be quite challenging to be diagnosed and no clear diagnostic flowchart has been established. Thus, a conjoined initiative on these two topics has been initiated by the European Society of Radiology (ESR), the European Association of Nuclear Medicine (EANM), the European Bone and Joint Infection Society (EBJIS), and the European Society of Clinical Microbiology and Infectious Diseases (ESCMID). The purpose of this work is to provide an overview on the two consensus documents on PBI and PJI that originated by the conjoined work of the ESR, EANM, and EBJIS (with ESCMID endorsement). METHODS AND RESULTS: After literature search, a list of 18 statements for PBI and 25 statements for PJI were drafted in consensus on the most debated diagnostic challenges on these two topics, with emphasis on imaging. CONCLUSIONS: Overall, white blood cell scintigraphy and magnetic resonance imaging have individually demonstrated the highest diagnostic performance over other imaging modalities for the diagnosis of PBI and PJI. However, the choice of which advanced diagnostic modality to use first depends on several factors, such as the benefit for the patient, local experience of imaging specialists, costs, and availability. Since robust, comparative studies among most tests do not exist, the proposed flowcharts are based not only on existing literature but also on the opinion of multiple experts involved on these topics. KEY POINTS: • For peripheral bone infection and prosthetic joint infection, white blood cell and magnetic resonance imaging have individually demonstrated the highest diagnostic performance over other imaging modalities. • Two evidence- and expert-based diagnostic flowcharts involving variable combination of laboratory tests, biopsy methods, and radiological and nuclear medicine imaging modalities are proposed by a multi-society expert panel. • Clinical application of these flowcharts depends on several factors, such as the benefit for the patient, local experience, costs, and availability.

A Rare Case of Chromoblastomycosis in a 12-year-old boy.

Chromoblastomycosis is a chronic fungal infection of the subcutaneous tissue. The infection usually results from a traumatic injury and inoculation of the microorganism by a specific group of dematiaceous fungi, resulting in the formation of verrucous plaques. The fungi produce sclerotic or medlar bodies (also called muriform bodies or sclerotic cells) seen on direct microscopic examination of skin smears. The disease is often found in adults due to trauma. We report a case of chromoblastomycosis in a 12-year-old child in whom the infection started when he was only 4 years old with secondary involvement of bones, cartilage, tongue and palatine tonsils. The child was not immunosuppressed.

Diagnostic challenges in pyogenic spinal infection: an expanded role for FDG-PET/CT.

In a preliminary investigation of FDG-PET/CT for assessment of therapy response of pyogenic spine infection, it was concluded that activity confined to the margins of a destroyed or degenerated joint with bone-on-bone contact represents nonseptic inflammation, regardless of the intensity of uptake. Only activity in bone, soft tissue, or within the epidural space represents active infection. The purpose of this investigation was to assess the performance of these pattern-based interpretation criteria in a series of problem cases of proven or suspected spine infection. Eighty-two FDG-PET/CTs were done for initial diagnosis because other imaging failed to provide a definitive diagnosis and 147 FDG-PET/CTs were done to assess treatment responses. Pattern-based interpretations were compared with the clinical diagnosis based on bacterial cultures and outcomes after cessation or withholding of antibiotic therapy. Pattern-based interpretation criteria achieved a sensitivity and specificity of 98 and 100%, respectively, for initial diagnosis and a specificity of 100% for assessment of treatment response. The same data was analyzed using intensity of activity as the primary factor. Sensitivity and specificity using the intensity-based criteria were 93 and 68%, respectively, for initial diagnosis, and the specificity of a negative interpretation for therapy response was 55%. Differences from pattern-based criteria are highly significant. Pattern-based criteria perform well in problem cases with equivocal MR and for treatment response because they correctly eliminate activity from nonspecific inflammation associated with destroyed joints with bone-on-bone contact. Response occurs within a timeframe that is useful for managing antibiotic therapy.

Mandibular Coronoidectomy Could Significantly Accelerate the Healing Process of Infratemporal Fossa Abscess.

PURPOSE: Management of an infratemporal fossa abscess (IFA), which is a specific form of severe and advanced deep fascial space infection (DFI), is based mainly on traditional methods. The purpose of this study was to investigate the role of mandibular coronoidectomy in accelerating IFA healing. PATIENTS AND METHODS: This research is a single-center retrospective study composed of 23 patients with IFA. The predictor variables were gender, age, diabetes, severity score, and mandibular coronoidectomy. The outcome variables included hospitalization time (HT) and irrigating time (IT). A comparison of treatment outcomes between the improved and traditional surgical interventions for IFA was performed. RESULTS: Compared with patients who did not receive mandibular coronoidectomy (NC group; HT, 17.54 ± 1.80 days; IT, 38.54 ± 3.73 days), patients who underwent mandibular coronoidectomy (AC group) had significantly decreased HT (7.20 ± 1.19 days) and IT (15.10 ± 1.27 days; P < .01). In addition, 4 patients (31%) in the NC group received reoperation for osteomyelitis, whereas no osteomyelitis and DFI recurrence occurred in the AC group. CONCLUSIONS: Mandibular coronoidectomy with extra intraoral drainage could considerably accelerate the healing process of IFAs and obviously decrease the reoperation rate for osteomyelitis.

Increased and normalized uptake of 18F-FDG in a case of bone periprosthetic infection treated by antibiotics.

We report the case of a 69 years old man with left hip prosthesis, who presented clinical, biochemical and imaging signs of periprosthetic infection treated with linezolid, an antibacterial agent of the oxazolidinone class. Two weeks after this treatment, a fluorine-18-fluoro-2-deoxy-d-glucose positron emission tomography/computed tomography (18F-FDG PET/CT) scan showed increased uptake in the skeleton and also increased uptake in several focal areas in the spine and near the prosthesis and the surgical wound on the left gluteus medius. Bone marrow biopsy was negative; meanwhile the antibiotic therapy, after four weeks of treatment was stopped due to red blood cells and platelets toxicity. Six weeks later, the patient developed high fever again and in order to revaluate the periprosthetic inflammation, he was resubmitted to 18F-FDG PET/CT which showed normal 18F-FDG uptake in the whole skeleton, including the prosthesis and the subcutaneous wound. Some focal areas of increased uptake in the lumbar spine were still detected. In the next 4 weeks the patient was under a "watch and wait" follow-up in a steady state. IN CONCLUSION: In the case we report, since we found no otnt focal areas in the lumbar spine where due to age-related bone deformities including some Schmorl's nodes. The inflammation in the bone prosthesis and the subcutaneous wound responded almost totally to the antibiotic treatment we applied.

Bone involvement by Sporothrix schenckii in an immunocompetent child.

Sporotrichosis in children is rare, and its osteoarticular form is very unusual. Disseminated forms are described mostly in immunocompromised patients. We report a case of a 5-year-old immunocompetent boy with multiple suppurated cutaneous lesions that progressed to polyarthritis of the hands and feet. Radiographic imaging demonstrated multifocal lytic lesions. Sporotrichosis was diagnosed through biopsy and culture. This article describes the radiographic appearance of a rare manifestation of this disease. In areas of high prevalence, the diagnosis of sporotrichosis should be taken into account, even in immunocompetent patients, when dactylitis with lytic lesions is present.

Fluorodeoxyglucose Positron Emission Tomography/Computed Tomography Interpretation Criteria for Assessment of Antibiotic Treatment Response in Pyogenic Spine Infection.

PURPOSE: The objective of the study was to determine if fluorodeoxyglucose positron emission tomography/computed tomography ((18)F-FDG PET/CT) can assess the response of patients with pyogenic spine infection to antibiotic treatment in a clinically useful time frame. METHODS: Twenty-eight patients with suspected pyogenic spine infection had baseline (18)F-FDG PET/CT. Patients with proven or probable infection were divided into good and poor responders to antibiotic therapy based on clinical criteria. These patients had a follow-up (18)F-FDG PET/CT 6-8 weeks later. RESULTS: Six of 28 patients were deemed negative for infection based on (18)F-FDG PET/CT. Two patients were excluded because of discrepancies in interpretation. Of the 20 patients deemed positive for infection, 13 had a pathogen isolated and all showed (18)F-FDG uptake in bone and/or soft tissue at baseline. Patients with a poor clinical response to treatment had persistent (18)F-FDG uptake in bone and/or soft tissue on follow-up. Patients with good clinical response had uptake confined to the margins of the destroyed disc. None of these patients had recurrent infection, even if antibiotics had already been discontinued at the time of the follow-up scan. CONCLUSIONS: (18)F-FDG uptake confined to the margins of a destroyed disc after antibiotic therapy of pyogenic spine infection must not be considered indicative of persistent infection and likely represents mechanically induced inflammation. (18)F-FDG uptake in bone or soft tissue does indicate active infection. Quantification of activity could not reliably differentiate patients with active infection from those without active infection and those who had had a successful response to therapy. The pattern of activity is critical to accurate interpretation.

Rapidly growing non-tuberculous mycobacterial infection in a renal transplant patient after alemtuzumab induction.

Renal transplant recipients are at high risk of developing opportunistic infections particularly in the first 6 months after transplantation. Organisms causing such infections include rapidly growing non-tuberculous mycobacteria (NTM). Lymphocytes have a central role in combating mycobacterial infections. The use of lymphocyte-depleting agents, such as alemtuzumab, in the renal transplant population has increased in recent years. A case of multifocal osteomyelitis caused by one of the NTM, Mycobacterium chelonae, in a renal transplant recipient, after alemtuzumab induction, is presented.

A large retrospective single-centre study to define the best image acquisition protocols and interpretation criteria for white blood cell scintigraphy with 99mTc-HMPAO-labelled leucocytes in musculoskeletal infections.

PURPOSE: The diagnosis of infection is often based on clinical, pathological and microbiological results. However, these investigations lack specificity. White blood cell (WBC) scintigraphy is considered the gold standard nuclear imaging technique for diagnosing infections in bone and soft tissues (except spondylodiscitis). However, image acquisition and interpretation criteria differ amongst centres throughout the world, leading to differences in reported results. The aim of this study was to identify the most accurate WBC scintigraphy acquisition and interpretation protocols for diagnosis of bone and soft tissue infections. METHODS: Included in this retrospective study were 297 patients with suspected bone or soft tissue infection who underwent WBC scintigraphy with (99m)Tc-HMPAO-labelled leucocytes between 2009 and 2012. Sensitivity, specificity, accuracy, and positive and negative predictive values of WBC scintigraphy were determined for two different dual time point acquisition protocols (fixed-time acquisition and time decay-corrected acquisition) and five image interpretation methods (visual and semiquantitative with four different reference regions of interest). Final diagnosis was based on pathological and microbiological reports, and when these were not available, on clinical follow-up of at least 6 months. RESULTS: The best acquisition protocol was 4 h and 20 - 24 h dual time-point acquisition with time decay-corrected acquisition. When using this acquisition protocol, visual qualitative interpretation led to a sensitivity of 85.1 %, a specificity of 97.1 %, a diagnostic accuracy of 94.5 %, a positive predictive value of 88.8 % and a negative predictive value of 95.9 %. For semiquantitative analysis, the best results were found when lesion-to-reference ratios were calculated with the contralateral side as the reference tissue, except for osteomyelitis and infected osteosynthesis, for which the contralateral bone marrow was found to be the best reference tissue. Results of the semiquantitative analyses per se were not better than for visual analysis. In the optimal analysis protocol, scans are first visually evaluated, and if this gives equivocal results, semiquantitative analysis is performed. This strategy resulted in an improved sensitivity of 97.9 %, a specificity of 91.8 % and a diagnostic accuracy of 93.1 %. CONCLUSION: WBC scintigraphy for bone and soft-tissue infection is best performed using a dual acquisition protocol at 4 h and at 20-24 h after injection, in which the acquisition time of the scans is corrected for decay. In most patients, visual analysis is sufficient and leads to high diagnostic accuracy. When interpretation by visual analysis is inconclusive, semiquantitative analysis adds accuracy. Based on our results, we propose a flow chart for analysing WBC scintigraphy in musculoskeletal infections.

A case of tertiary neurosyphilis presenting with moth-eaten bone lesions.

Syphilis, the "great imitator," with regard to skin diseases, is a chronic systemic infectious disease with a clinical course that waxes and wanes. The incidence of tertiary syphilis had decreased drastically these decades. We report a case of tertiary neurosyphilis presenting with moth-eaten bone lesions of the lower extremities. To the best of our knowledge, we have not seen such reports.

Superimposed infection in mandibular osteoradionecrosis: diagnosis and outcomes.

BACKGROUND: Radiation therapy can result in osteoradionecrosis (ORN) and mucosal ulceration predisposing to infection. METHODS: Fourteen patients presenting with infectious sequelae related to mandibular ORN were retrospectively reviewed. RESULTS: In most patients, infection followed diagnosis of ORN; but in 4 patients, ORN was not diagnosed until after the time of infection and imaging. An early imaging finding of ORN was lingual cortical defects near the last molar. Pain followed by erythema, purulent drainage, and subperiosteal abscess by imaging were the most common signs of infection. In most patients, conservative management eventually failed and segmental mandibulectomies were required. CONCLUSIONS: Soft tissue infection with characteristic bone findings such as subperiosteal abscess and cortical bone erosions helps to distinguish infected ORN from recurrent tumor or sterile ORN. In patients previously treated with radiation who present with infection, pain or an avid PET scan with bone involvement, the mandible should be scrutinized.

Treatment of infected tibial nonunion with bone defect using central bone grafting technique.

Treatment of infected tibial nonunion with bone defect represents a challenge for every orthopaedic surgeon. Various methods of treatment have been described for nonunions with infection, bone loss or both. One of them is the central bone grafting technique, which is a safe and effective treatment for nonunions of the tibia. The technique involves placement of autogenous cancellous bone from the iliac crest on the anterior surface of the interosseous membrane with the aim of creating a tibiofibular synostosis. We present the results of uncontrolled, retrospective and continuous series of ten patients treated by a central bone grafting technique for infected tibial nonunion with bone loss. Mean follow-up period was 12 (10-15) years. Most injuries were a result of war injuries. Clinically and radiologically confirmed bony healing with total consolidation of the graft was achieved in all patients within a period of 10-12 months without further bone grafting. The newly-formed bone mass was able to fulfil the mechanical and functional demands of everyday life activities. Once again, the central bone grafting technique has shown to be a safe, reliable and effective method of treatment for infected tibial nonunion with bone defect.