Role of diffusion-weighted imaging for detecting bone marrow infiltration in skull in children with acute lymphoblastic leukemia

Radiomics model and deep learning model based on T1WI image for acute lymphoblastic leukemia identification

AIM

This study aimed to develop highly precise radiomics and deep learning models to accurately detect acute lymphoblastic leukemia (ALL) using a T1WI image.

MATERIALS AND METHODS

A total of 604 brain magnetic resonance data of ALL group and normal children (NC) group. Two radiologists independently retrieved radiomics features after manually delineating the area of interest along the clivus at the median sagittal position of T1WI. According to the 9:1 ratio, all samples were randomly divided into the training cohort and the testing cohort. support vector machine was then used to classify the radiomics model using the features that had a correlation coefficient of greater than 0.99 in the training cohort. The Efficientnet-B3 network model received the training set images to create a deep learning model. The sensitivity, specificity, and area under the ROC curve were calculated in order to evaluate the diagnostic efficacy of the different models after the validation of two aforementioned models in the testing cohort.

RESULTS

The deep learning model had a higher AUC value of 0.981 than the radiomics model's value of 0.962 in the testing cohort. Delong's test showed no statistical difference between the two models (P>0.05). The accuracy/sensitivity/specificity/negative predictive value/positive predictive value achieved 0.9180/0.9565/0.8947/0.9714/0.8462 for the radiomics model and 0.9344/0.8696/0.9737/0.9250/0.9524 for deep learning model.

CONCLUSIONS

The deep learning and radiomics models showed high AUC values in the training and test cohorts. They also exhibited good diagnostic efficacy for predicting ALL.

Magnetic resonance imaging study of normal cranial bone marrow conversion at high altitude

BACKGROUND

To use conventional magnetic resonance imaging (MRI) and diffusion-weighted imaging (DWI) to investigate the effects of long-term hypoxia on cranial bone marrow conversion in healthy people at high altitudes.

METHODS

A total of 1,130 individuals were selected from altitudinal areas of 2,000-3,000, 3,100-4,000, and >4,100 m. Each altitude range was divided into 5 age groups: 0-5, 6-14, 15-29, 30-49, and ≥50 years. Firstly, cranial bone marrow typing of the participants in each altitude range was performed on sagittal T1-weighted images (T1WI) according to the average diploe thickness and signal intensity of the normal skull, and the relationship between bone marrow conversion and age was analyzed. Secondly, the apparent diffusion coefficient (ADC) values of the frontal bone, parietal bone, occipital bone, and temporal bone were measured in the DWI post-processing workstation and statistical methods were used to analyze whether different altitudinal gradients and long-term hypoxic environment had any effect on cranial bone marrow conversion.

RESULTS

There was a positive correlation between bone marrow type and age in the healthy populations at all 3 levels of altitude (P<0.05). The average thickness of the cranial diploe also positively correlated with age (P<0.05); in the age ranges of 30-49 and ≥50 years, the ADC values of the occipital and temporal bone marrow positively correlated with increasing altitude (P<0.05).

CONCLUSIONS

The cranial bone marrow of normal people at high altitudes changes from Type I to Type IV with increasing age and under the influence of long-term chronic hypoxia. The bone marrow of the occipital and temporal bones of healthy people aged 30-49 and ≥50 years showed erythromedularization during the process of Type III and IV bone marrow conversion.

Targeted blockade of immune mechanisms inhibit B precursor acute lymphoblastic leukemia cell invasion of the central nervous system

Acute lymphoblastic leukemia (ALL) dissemination to the central nervous system (CNS) is a challenging clinical problem whose underlying mechanisms are poorly understood. Here, we show that primary human ALL samples injected into the femora of immunodeficient mice migrate to the skull and vertebral bone marrow and provoke bone lesions that enable passage into the subarachnoid space. Treatment of leukemia xenografted mice with a biologic antagonist of receptor activator of nuclear factor κB ligand (RANKL) blocks this entry route. In addition to erosion of cranial and vertebral bone, samples from individuals with B-ALL also penetrate the blood-cerebrospinal fluid barrier of recipient mice. Co-administration of C-X-C chemokine receptor 4 (CXCR4) and RANKL antagonists attenuate both identified routes of entry. Our findings suggest that targeted RANKL and CXCR4 pathway inhibitors could attenuate routes of leukemia blast CNS invasion and provide benefit for B-ALL-affected individuals.

[Skeletal manifestations of systemic hematologic disorders]

BACKGROUND

Bone marrow consists of connective tissue and stem cells, which generate blood cells. This includes erythropoiesis, leukopoiesis and thrombopoiesis. Thus, hematologic disorders first affect the bone marrow and secondarily the blood.

METHODS

Bone marrow changes can be sensitively detected using magnetic resonance imaging (MRI) and often represent the initial manifestation of the underlying disease. With longer duration of disease, changes can also be found on X‑ray or computed tomography (CT).

RESULTS

The findings on MRI and X‑ray/CT are often nonspecific and can only be interpreted in the context of clinical information.

CONCLUSION

In the following article, we provide a brief overview of the clinical manifestations and imaging changes to be expected in leukemia, anemia, and chronic myeloproliferative disorders.

Imaging Spectrum of Calvarial Abnormalities

Calvarial abnormalities are usually discovered incidentally on radiologic studies or less commonly manifest with symptoms. This narrative review describes the imaging spectrum of the abnormal calvaria. The extent, multiplicity, and other imaging features of calvarial abnormalities can be combined with the clinical information to establish a final diagnosis or at least narrow the differential considerations. Prior trauma (congenital depression, leptomeningeal cysts, posttraumatic osteolysis), surgical intervention (flap osteonecrosis and burr holes), infection, and inflammatory processes (sarcoidosis) can result in focal bone loss, which may also be seen with idiopathic disorders without (bilateral parietal thinning and Gorham disease) or with (Parry-Romberg syndrome) atrophy of the overlying soft tissues. Anatomic variants (arachnoid granulations, venous lakes, parietal foramina) and certain congenital lesions (epidermoid and dermoid cysts, atretic encephalocele, sinus pericranii, and aplasia cutis congenita) manifest as solitary lytic lesions. Other congenital entities (lacunar skull and dysplasia) display a diffuse pattern of skull involvement. Several benign and malignant primary bone tumors involve the calvaria and manifest as lytic, sclerotic, mixed lytic and sclerotic, or thinning lesions, whereas multifocal disease is mainly due to hematologic or secondary malignancies. Metabolic disorders such as rickets, hyperparathyroidism, renal osteodystrophy, acromegaly, and Paget disease involve the calvaria in a more diffuse pattern. Online supplemental material is available for this article. ^©RSNA, 2021.

Total diffusion volume in MRI vs. total lesion glycolysis in PET/CT for tumor volume evaluation of multiple myeloma

OBJECTIVE

This study compared the tumor burden and prognostic impact of total diffusion volume (tDV) and total lesion glycolysis (TLG) in the same patients with newly diagnosed multiple myeloma (NDMM) simultaneously. We also examined the relationship between these imaging tumor volumes (TVs) and plasma cell (PC) TV in bone marrow (BM) specimens.

METHODS

We retrospectively reviewed the data of 63 patients with newly diagnosed multiple myeloma (NDMM) from April 2016 to March 2018. tDV was calculated from whole-body diffusion-weighted imaging and TLG was calculated from the average standard uptake value and the metabolic tumor volume, respectively. Cellularity of BM hematopoietic tissue and the percentage of BM PCs were used as a reference of PC volume in the BM.

RESULTS

The Spearman correlation coefficient between tDV and TLG was moderate (ɤs = 0.588, p < 0.001) when PET false-negative patients were excluded. There were positive correlations between the BM plasma cell volume (BMPCV) and the imaging TVs (ɤs = 0.505, vs. tDV; and 0.464, vs. TLG). Patients with high tDV and high TLG, as determined by the receiver operating characteristic curve, had worse survival; moreover, patients with both high tDV and high TLG showed the worst prognosis (median progression-free and overall survival: 13.2 and 28.9 months, respectively).

CONCLUSIONS

Although tDV and TLG each reflected the total TV, in several cases, tDV and TLG were discrepant due to the biological features of each MM. It is important to use both modalities for complementary assessment of total tumor burden and biological characteristics in MM.

KEY POINTS

• Total diffusion volume (tDV) and total lesion glycolysis (TLG) reflect the total tumor volume and have prognostic value in patients with multiple myeloma (MM). • tDV and TLG could assess MM from different biological perspectives and should be considered for each patient individually.

Role of magnetic resonance imaging in characterization of central nervous system lesions in pediatric patients with leukemia and post-treatment complications

Background

Leukemia is one of the most common fatal diseases in pediatric oncology. Recently, advances in drug therapy have improved the prognosis of acute leukemia with event-free survival of up to 60%; however, complications and adverse effects of the disease and anti-leukemic treatment have also increased. The CNS complications of leukemia can be classified into those that developed directly or indirectly from the underlying leukemic process and those that can be related to antileukemic therapy. MRI had improved early detection of CNS complications and proper management. The study aims to characterize the MRI findings caused by the leukemic involvement of CNS structures and treatment-associated CNS complications and assess its value in early management and avoidance of long-term side effects.

Results

The patient’s age ranged from 2 to 18 years with different types of leukemia classified regarding the time of presentation as pretreatment, during treatment phases, and post-treatment. Different MRI abnormalities were recorded and clinically correlated.

Conclusion

The neurological complications of leukemia have common presenting symptoms but varying imaging abnormalities. To reach the correct diagnosis, the presenting signs, symptoms, and laboratory data must be considered along with the radiologic findings. A diagnostic algorithm using conventional, post-contrast MRI, MR venography, along with diffusion-weighted MRI was of great value in early detection and differentiation of different CNS lesions detected in pediatric patients with leukemia and post-treatment CNS complications.

Risk Factors for Relapse of Childhood B Cell Acute Lymphoblastic Leukemia

Background

B cell acute lymphoblastic leukemia (B-ALL) is the most common type of ALL. This study aimed to explore risk factors for relapse of childhood B-ALL.

Material/Methods

Total of 102 pediatric B-ALL patients were included in this study. B-ALL patients were divided into a relapse group and a non-relapse group. Chemotherapy-induced agranulocytosis time, fusion gene, and minimal residual disease (MRD) were assessed. White blood cell (WBC) count in peripheral blood and risk stratification were evaluated in newly-diagnosed patients. Kaplan-Meier plots were used to evaluate the correlation between risk factors and relapse rates. Multivariate analysis was performed with Cox proportional hazard model to estimate relative risk (RR), 95% confidence interval (95% CI), and hazard ratio (HR). Finally, 99 cases of B-ALL were included in this study.

Results

There were significant differences between the relapse group and the non-relapse group in age (p=0.004), chemotherapy-induced agranulocytopenia (p=0.001), WBC count in peripheral blood of newly diagnosed patients (p=0.016), risk stratification (p=0.000), and MRD at 12^th week (p=0.007). Age over 10 years, high-risk stratification, long period of agranulocytopenia, higher WBC counts, and MRD more than 10⁻⁴ were correlated with higher B-ALL relapse rate (p<0.05). Multivariate analysis showed significantly higher relapse rates for age ≥10 years, high-risk stratification, and MRD at 12^th week >10⁻⁴, with RR (95% CI) of 4.001 (1.005–15.930), 4.964 (1.050–23.456), and 4.646 (1.383–15.614), respectively.

Conclusions

Agranulocytopenia ≤7 days, peripheral blood WBC >100×10⁹/L, and MRD at 33^rd day >10⁻⁴ were associated with B-ALL relapse. Age ≥10 years, high-risk stratification, and MRD at 12^th week >10⁻⁴ were independent risk factors for relapse.

Diffusion MR Imaging in the Head and Neck: Principles and Applications

Diffusion imaging is a functional MR imaging tool that creates tissue contrast representative of the random, microscopic translational motion of water molecules within human body tissues. Long considered a cornerstone MR imaging sequence for brain imaging, diffusion-weighted imaging (DWI) increasingly is used for head and neck imaging. This review reports the current state of diffusion techniques for head and neck imaging, including conventional DWI, DWI trace with apparent diffusion coefficient map, diffusion tensor imaging, intravoxel incoherent motion, and diffusion kurtosis imaging. This article describes background physics, reports supportive evidence and potential pitfalls, highlights technical advances, and details practical clinical applications.

Integrin-Mediated Adhesion and Chemoresistance of Acute Lymphoblastic Leukemia Cells Residing in the Bone Marrow or the Central Nervous System

Acute Lymphoblastic Leukemia (ALL) is the most common cancer in childhood. Despite a significantly improved prognosis over the last decade with a 5-years survival rate of ~90%, treatment-related morbidity remains substantial and relapse occurs in 10–15% of patients (1). The most common site of relapse is the bone marrow, but early colonization and subsequent reoccurrence of the disease in the central nervous system (CNS) also occurs. Integrins are a family of cell surface molecules with a longstanding history in cancer cell adherence, migration and metastasis. In chronic lymphoblastic leukemia (CLL), the VLA-4 integrin has been acknowledged as a prognostic marker and mounting evidence indicates that this and other integrins may also play a role in acute leukemia, including ALL. Importantly, integrins engage in anti-apoptotic signaling when binding extracellular molecules that are enriched in the bone marrow and CNS microenvironments. Here, we review the current evidence for a role of integrins in the adherence of ALL cells within the bone marrow and their colonization of the CNS, with particular emphasis on mechanisms adding to cancer cell survival and chemoresistance.

Benign and malignant skull-involved lesions: discriminative value of conventional CT and MRI combined with diffusion-weighted MRI

BACKGROUND

Little is known about the value of computed tomography (CT) and magnetic resonance imaging (MRI) combined with diffusion-weighted imaging (DWI) in distinguishing malignant from benign skull-involved lesions.

PURPOSE

To evaluate the discriminative value of DWI combined with conventional CT and MRI for differentiating between benign and malignant skull-involved lesions.

MATERIAL AND METHODS

CT and MRI findings of 58 patients with pathologically proven skull-involved lesions (43 benign and 15 malignant) were retrospectively reviewed. Conventional CT and MRI characteristics and apparent diffusion coefficient (ADC) value of the two groups were evaluated and compared. Multivariate logistic regression and receiver operating characteristic (ROC) curve analyses were performed to assess the differential performance of each parameter separately and together.

RESULTS

The presence of cortical defects or break-through and ill-defined margins were associated with malignant skull-involved lesions (both P < 0.05). Malignant skull-involved lesions demonstrated a significantly lower ADC (P = 0.016) than benign lesions. ROC curve analyses indicated that a combination of CT, MRI, and DWI with an ADC ≤ 0.703 × 10^-3 mm²/s showed optimal sensitivity, while DWI along showed optimal specificity of 88.4% in differentiating between benign and malignant skull-involved lesions.

CONCLUSION

The combination of CT, MRI, and DWI can help to differentiate malignant from benign skull-involved lesions. CT + MRI + DWI offers optimal sensitivity, while DWI offers optimal specificity.

MRI and PET/MRI in hematologic malignancies

The role of MRI differs considerably between the three main groups of hematological malignancies: lymphoma, leukemia, and myeloma. In myeloma, whole‐body MRI (WB‐MRI) is recognized as a highly sensitive test for the assessment of myeloma, and is also endorsed by clinical guidelines, especially for detection and staging. In lymphoma, WB‐MRI is presently not recommended, and merely serves as an alternative technique to the current standard imaging test, [¹⁸F]FDG‐PET/CT, especially in pediatric patients. Even for lymphomas with variable FDG avidity, such as extranodal mucosa‐associated lymphoid tissue lymphoma (MALT), contrast‐enhanced computed tomography (CT), but not WB‐MRI, is presently recommended, despite the high sensitivity of diffusion‐weighted MRI and its ability to capture treatment response that has been reported in the literature. In leukemia, neither MRI nor any other cross‐sectional imaging test (including positron emission tomography [PET]) is currently recommended outside of clinical trials. This review article discusses current clinical applications as well as the main research topics for MRI, as well as PET/MRI, in the field of hematological malignancies, with a focus on functional MRI techniques such as diffusion‐weighted imaging and dynamic contrast‐enhanced MRI, on the one hand, and novel, non‐FDG PET imaging probes such as the CXCR4 radiotracer [⁶⁸Ga]Ga‐Pentixafor and the amino acid radiotracer [¹¹C]methionine, on the other hand.

Level of Evidence: 5

Technical Efficacy Stage: 3

J. Magn. Reson. Imaging 2020;51:1325–1335.

Role of diffusion-weighted MRI in differentiating benign from malignant bone tumors

Objective:

To evaluate the role of diffusion-weighted MRI in differentiating benign from malignant primary bone tumors. To know the sensitivity and specificity of diffusion weighted MRI and calculating apparent diffusion coefficient (ADC) cutoff in differentiating benign from malignant primary bone tumors.

Methods and materials :

This is a prospective observational study of 50 patients, who were clinically or radiologically suspected with primary bone tumor and referred to the Department of Radiodiagnosis, for radiography or for MRI. These patients underwent routine MRI sequences including diffusion-weighted MRI with b-values of 0, 500 and 1000, followed by pathological examination supplemented by immunohistochemistry (wherever necessary). Hematological malignancies, recently biopsied cases and recurrent cases were excluded from the study.

Results:

Out of 50 patients with suspected bone tumors, 15 were benign (and tumor like lesions) and 35 were malignant primary bone tumors. The most common age group involved for both benign and malignant primary bone tumors was 11–20 years (23 cases—46%). In our study, total number of affected males were 27 (54%) and total number of affected females were 23 (46%) with M:F ratio of 1.17:1. In this study 72% lesions had appendicular bone involvement and 28% had axial bone involvement. 94.3% of malignant lesions showed restriction on diffusion-weighted imaging (DWI) and in 80 % of benign lesions restriction was absent on DWI which was statistically significant. Mean ADC levels in malignant lesions was 1.092 ± 0.497 and in benign lesions was 1.62 ± 0.596 which was statistically significant. Chondrosarcoma had highest ADC and Ewing’s sarcoma had lowest ADC values in malignant lesions. Chondroblastoma had highest ADC and Osteomyelitis had lowest ADC values in benign lesions. ADC value of 1.31 had highest sensitivity and specificity to differentiate between benign and malignant lesions.

Conclusion:

DWI is helpful in differentiating malignant from benign bone tumors and tumor like lesions with diffusion restriction favoring malignancy. Inspite of some overlap, ADC values of benign and malignant bone tumors are different and measurement of ADC values improves the accuracy of the diagnosis of bone tumors and tumor like lesions. Calculation of ADC may also be used as baseline reference to assess response to treatment in future or for follow up.

Advances in knowledge:

DWI imaging (and ADC values) has been extensively used in neuroimaging. Extension of this application to musculoskeletal–oncologic imaging is not so well studied. Apart from differentiating benign from malignant lesions which is the main focus of this study, assessment of response to treatment by ADC values may be possible in near future.

Questionable correlation of the apparent diffusion coefficient with the histological grade and microvascular invasion in small hepatocellular carcinoma

AIM

To evaluate the correlation between the apparent diffusion coefficient (ADC) and various histopathological parameters in small hepatocellular carcinomas (HCCs).

MATERIALS AND METHODS

In 143 surgically resected small HCCs, the mean and minimum ADC values, tumour-to-liver ADC ratio, and normalised ADC (ADC of the HCC/ADC of the spleen) were correlated to the tumour grade, microvascular invasion (MVI), cellularity, fatty change, degree of fibrosis, and lymphocytic infiltration using linear regression analysis, the Wilcoxon rank sum test, or Spearman's rank correlation.

RESULTS

No significant correlation was found between the ADC parameters and tumour grade. In the univariate analysis, the ADC ratio of the tumour was significantly correlated with MVI as well as the degree of fibrosis and lymphocyte infiltration of the HCC (p=0.017, 0.042, and 0.002, respectively). The ADC of the tumour was significantly correlated with the degree of lymphocyte infiltration of the HCC (p=0.049). In the multivariate analysis, the ADC ratio of the tumour was an independent parameter for MVI and the degree of lymphocyte infiltration of the HCC (p=0.034 and <0.001, respectively), and the ADC of the tumour was an independent parameter for the degree of lymphocyte infiltration of the HCC (p=0.009). There was no significant correlation between the other ADCs and pathological tumour parameters.

CONCLUSION

The tumour grade of small HCCs was not correlated with ADC parameters. The tumour-to-liver ADC ratio was a significant independent parameter for the degree of lymphocyte infiltration and MVI of small HCCs.

Craniofacial Manifestations of Systemic Disorders: CT and MR Imaging Findings and Imaging Approach

Many systemic diseases or conditions can affect the maxillofacial bones; however, they are often overlooked or incidentally found at routine brain or head and neck imaging performed for other reasons. Early identification of some conditions may significantly affect patient care and alter outcomes. Early recognition of nonneoplastic hematologic disorders, such as thalassemia and sickle cell disease, may help initiate earlier treatment and prevent serious complications. The management of neoplastic diseases such as lymphoma, leukemia, or Langerhans cell histiocytosis may be different if diagnosed early, and metastases to the maxillofacial bones may be the first manifestation of an otherwise occult neoplasm. Endocrinologic and metabolic disorders also may manifest with maxillofacial conditions. Earlier recognition of osteoporosis may alter treatment and prevent complications such as insufficiency fractures, and identification of acromegaly may lead to surgical treatment if there is an underlying growth hormone-producing adenoma. Bone dysplasias sometimes are associated with skull base foraminal narrowing and subsequent involvement of the cranial nerves. Inflammatory processes such as rheumatoid arthritis and sarcoidosis may affect the maxillofacial bones, skull base, and temporomandibular joints. Radiologists should be familiar with the maxillofacial computed tomographic and magnetic resonance imaging findings of common systemic disorders because these may be the first manifestations of an otherwise unrevealed systemic process with potential for serious complications. Online supplemental material is available for this article. ^©RSNA, 2018.