2.5D and 3D segmentation of brain metastases with deep learning on multinational MRI data.

Introduction: Management of patients with brain metastases is often based on manual lesion detection and segmentation by an expert reader. This is a time- and labor-intensive process, and to that end, this work proposes an end-to-end deep learning segmentation network for a varying number of available MRI available sequences. Methods: We adapt and evaluate a 2.5D and a 3D convolution neural network trained and tested on a retrospective multinational study from two independent centers, in addition, nnU-Net was adapted as a comparative benchmark. Segmentation and detection performance was evaluated by: (1) the dice similarity coefficient, (2) a per-metastases and the average detection sensitivity, and (3) the number of false positives. Results: The 2.5D and 3D models achieved similar results, albeit the 2.5D model had better detection rate, whereas the 3D model had fewer false positive predictions, and nnU-Net had fewest false positives, but with the lowest detection rate. On MRI data from center 1, the 2.5D, 3D, and nnU-Net detected 79%, 71%, and 65% of all metastases; had an average per patient sensitivity of 0.88, 0.84, and 0.76; and had on average 6.2, 3.2, and 1.7 false positive predictions per patient, respectively. For center 2, the 2.5D, 3D, and nnU-Net detected 88%, 86%, and 78% of all metastases; had an average per patient sensitivity of 0.92, 0.91, and 0.85; and had on average 1.0, 0.4, and 0.1 false positive predictions per patient, respectively. Discussion/Conclusion: Our results show that deep learning can yield highly accurate segmentations of brain metastases with few false positives in multinational data, but the accuracy degrades for metastases with an area smaller than 0.4 cm2.

MRI pulse sequence integration for deep-learning-based brain metastases segmentation.

PURPOSE: Magnetic resonance (MR) imaging is an essential diagnostic tool in clinical medicine. Recently, a variety of deep-learning methods have been applied to segmentation tasks in medical images, with promising results for computer-aided diagnosis. For MR images, effectively integrating different pulse sequences is important to optimize performance. However, the best way to integrate different pulse sequences remains unclear. In addition, networks trained with a certain subset of pulse sequences as input are unable to perform when given a subset of those pulse sequences. In this study, we evaluate multiple architectural features and characterize their effects in the task of metastasis segmentation while creating a method to robustly train a network to be able to work given any strict subset of the pulse sequences available during training. METHODS: We use a 2.5D DeepLabv3 segmentation network to segment metastases lesions on brain MR's with four pulse sequence inputs. To study how we can best integrate MR pulse sequences for this task, we consider (1) different pulse sequence integration schemas, combining our features at early, middle, and late points within a deep network, (2) different modes of weight sharing for parallel network branches, and (3) a novel integration level dropout layer, which will allow the networks to be robust to performing inference on input with only a subset of pulse sequences available at the training. RESULTS: We find that levels of integration and modes of weight sharing that favor low variance work best in our regime of small amounts of training data (n = 100). By adding an input-level dropout layer, we could preserve the overall performance of these networks while allowing for inference on inputs with missing pulse sequences. We illustrate not only the generalizability of the network but also the utility of this robustness when applying the trained model to data from a different center, which does not use the same pulse sequences. Finally, we apply network visualization methods to better understand which input features are most important for network performance. CONCLUSIONS: Together, these results provide a framework for building networks with enhanced robustness to missing data while maintaining comparable performance in medical imaging applications.

Handling missing MRI sequences in deep learning segmentation of brain metastases: a multicenter study.

The purpose of this study was to assess the clinical value of a deep learning (DL) model for automatic detection and segmentation of brain metastases, in which a neural network is trained on four distinct MRI sequences using an input-level dropout layer, thus simulating the scenario of missing MRI sequences by training on the full set and all possible subsets of the input data. This retrospective, multicenter study, evaluated 165 patients with brain metastases. The proposed input-level dropout (ILD) model was trained on multisequence MRI from 100 patients and validated/tested on 10/55 patients, in which the test set was missing one of the four MRI sequences used for training. The segmentation results were compared with the performance of a state-of-the-art DeepLab V3 model. The MR sequences in the training set included pre-gadolinium and post-gadolinium (Gd) T1-weighted 3D fast spin echo, post-Gd T1-weighted inversion recovery (IR) prepped fast spoiled gradient echo, and 3D fluid attenuated inversion recovery (FLAIR), whereas the test set did not include the IR prepped image-series. The ground truth segmentations were established by experienced neuroradiologists. The results were evaluated using precision, recall, Intersection over union (IoU)-score and Dice score, and receiver operating characteristics (ROC) curve statistics, while the Wilcoxon rank sum test was used to compare the performance of the two neural networks. The area under the ROC curve (AUC), averaged across all test cases, was 0.989 ± 0.029 for the ILD-model and 0.989 ± 0.023 for the DeepLab V3 model (p = 0.62). The ILD-model showed a significantly higher Dice score (0.795 ± 0.104 vs. 0.774 ± 0.104, p = 0.017), and IoU-score (0.561 ± 0.225 vs. 0.492 ± 0.186, p < 0.001) compared to the DeepLab V3 model, and a significantly lower average false positive rate of 3.6/patient vs. 7.0/patient (p < 0.001) using a 10 mm3 lesion-size limit. The ILD-model, trained on all possible combinations of four MRI sequences, may facilitate accurate detection and segmentation of brain metastases on a multicenter basis, even when the test cohort is missing input MRI sequences.

Responses in the diffusivity and vascular function of the irradiated normal brain are seen up until 18 months following SRS of brain metastases.

BACKGROUND: MRI may provide insights into longitudinal responses in the diffusivity and vascular function of the irradiated normal-appearing brain following stereotactic radiosurgery (SRS) of brain metastases. METHODS: Forty patients with brain metastases from non-small cell lung cancer (N = 26) and malignant melanoma (N = 14) received SRS (15-25 Gy). Longitudinal MRI was performed pre-SRS and at 3, 6, 9, 12, and 18 months post-SRS. Measures of tissue diffusivity and vascularity were assessed by diffusion-weighted and perfusion MRI, respectively. All maps were normalized to white matter receiving less than 1 Gy. Longitudinal responses were assessed in normal-appearing brain, excluding tumor and edema, in the LowDose (1-10 Gy) and HighDose (>10 Gy) regions. The Eastern Cooperative Oncology Group (ECOG) performance status was recorded pre-SRS. RESULTS: Following SRS, the diffusivity in the LowDose region increased continuously for 1 year (105.1% ± 6.2%; P < .001), before reversing toward pre-SRS levels at 18 months. Transient reductions in microvascular cerebral blood volume (P < .05), blood flow (P < .05), and vessel densities (P < .05) were observed in LowDose at 6-9 months post-SRS. Correspondingly, vessel calibers in LowDose transiently increased at 3-9 months (P < .01). The responses in HighDose displayed similar trends as in LowDose, but with larger interpatient variations. Vascular responses followed pre-SRS ECOG status. CONCLUSIONS: Our results imply that even low doses of radiation to normal-appearing brain following cerebral SRS induce increased diffusivity and reduced vascular function for up until 18 months. In particular, the vascular responses indicate the reduced ability of the normal-appearing brain tissue to form new capillaries. Assessing the potential long-term neurologic effects of SRS on the normal-appearing brain is warranted.

Soluble AXL as a marker of disease progression and survival in melanoma.

Receptor tyrosine kinase AXL is a one-pass transmembrane protein upregulated in cancers and associated with lower survival and therapy resistance. AXL can be cleaved by the A Disintegrin and Metalloproteinases (ADAM)10 and ADAM17, yielding a soluble version of the protein. Elevated soluble AXL (sAXL) has been reported to be associated with disease progression in hepatocellular carcinoma, renal cancer, neurofibromatosis type 1 and inflammatory diseases. In the present work, we analyzed sAXL levels in blood from melanoma patients and showed that sAXL increases with disease progression. Additionally, increased sAXL levels were found correlated with shorter two-year survival in stage IV patients treated with ipilimumab. Furthermore, we showed that sAXL levels were related to the percentage of cells expressing AXL in resected melanoma lymph node metastases. This finding was verified in vitro, where sAXL levels in the cell media corresponded to AXL expression in the cells. AXL inhibition using the small-molecular inhibitor BGB324 reduced sAXL levels, while the cellular expression was elevated through increased protein stability. Our findings signify that quantification of sAXL blood levels is a simple and easily assessable method to determine cellular AXL levels and should be further evaluated for its use as a biomarker of disease progression and treatment response.

Brain metastases with poor vascular function are susceptible to pseudoprogression after stereotactic radiation surgery.

PURPOSE: This study aimed to investigate the hemodynamic status of cerebral metastases prior to and after stereotactic radiation surgery (SRS) and to identify the vascular characteristics that are associated with the development of pseudoprogression from radiation-induced damage with and without a radionecrotic component. METHODS AND MATERIALS: Twenty-four patients with 29 metastases from non-small cell lung cancer or malignant melanoma received SRS with dose of 15 Gy to 25 Gy. Magnetic resonance imaging (MRI) scans were acquired prior to SRS, every 3 months during the first year after SRS, and every 6 months thereafter. On the basis of the follow-up MRI scans or histology after SRS, metastases were classified as having response, tumor progression, or pseudoprogression. Advanced perfusion MRI enabled the estimation of vascular status in tumor regions including fractions of abnormal vessel architecture, underperfused tissue, and vessel pruning. RESULTS: Prior to SRS, metastases that later developed pseudoprogression had a distinct poor vascular function in the peritumoral zone compared with responding metastases (P < .05; number of metastases = 15). In addition, differences were found between the peritumoral zone of pseudoprogressing metastases and normal-appearing brain tissue (P < .05). In contrast, for responding metastases, no differences in vascular status between peritumoral and normal-appearing brain tissue were observed. The dysfunctional peritumoral vasculature persisted in pseudoprogressing metastases after SRS. CONCLUSIONS: Our results suggest that the vascular status of peritumoral tissue prior to SRS plays a defining role in the development of pseudoprogression and that advanced perfusion MRI may provide new insights into patients' susceptibility to radiation-induced effects.

Accrual to a randomised trial of adjuvant whole brain radiotherapy for treatment of melanoma brain metastases is feasible.

BACKGROUND: Brain metastases (BMs) are common in melanoma patients. Adjuvant whole brain radiotherapy (WBRT) following local treatment of intracranial melanoma metastases with neurosurgery and/or stereotactic radiosurgery is controversial. A randomised trial is needed. However, accrual to WBRT trials has been problematic. A pilot study by Australia and New Zealand Melanoma Trials Group (ANZMTG) was conducted to see if accrual was feasible. METHODS: Sites canvassed for interest included those who treat melanoma patients, had a proven accrual in previous melanoma trials and who had the relevant infrastructure support. Feasibility forecasts from interested sites were sought. These were compared to the patient numbers documented in the research contracts. A target accrual of 60 patients in 2 years was set. Funding was sought for the pilot study. Basic demographics of the pilot study cohort were collected. RESULTS: The first centre opened December 2008; the first patient was randomised in April 2009. The pilot accruing period concluded in September, 2011. In 30 months, 54 patients from 10 of a total of 17 activated sites in Australia (39, 72%) and in Norway (15, 28%) had been accrued. Feasibility forecasts predicted 133 trial eligible patients per year (including 108 Australian + 25 International patients). Site estimates generally overestimated accrual with 4 of 17 active sites estimating within 50% of target numbers. Sites with patient estimates calculated from records were more accurate than those estimated from memory. The overall recruitment target was lower in the research contracts when compared to the feasibility evaluation. Basic demographics of the pilot study revealed 62% of patients were males; 58% had a single metastasis, 28% had two and 14% had three metastases. 12-month overall survival was 50%. CONCLUSIONS: Despite only 54 patients and not the full 60 patient target being accrued in two years the Trial Management Committee and Data Safely Monitoring Committee approved the continuation of the pilot study to the main trial. On the basis of this successful pilot study, funding was achieved for the full study. 143 patients of a target of 200 have been randomised by June 2014.

[Treatment of metastatic malignant melanoma localized to an extremity]. / Behandling av melanommetastaser på ekstremitetene.

BACKGROUND: At the Norwegian Radium Hospital, most patients who were treated for metastases (localized to an extremity) from malignant melanoma in the period 1977-99 underwent intra-arterial chemotherapy combined with radiotherapy. We have evaluated the effects of this treatment, which has now been replaced by isolated limb perfusion (with melphalan and tumor necrosis factor). MATERIAL AND METHODS: Medical records were reviewed for patients with metastatic malignant melanoma (localized to an extremity) who had been treated at the Norwegian Radium Hospital with intra-arterial chemotherapy (5-[3,3 dimethyl-1-triazeno]-imidazole-4-carboxamide [DTIC]) in the period 1977-99. RESULTS: 36 patients had received such treatment; in 30 of these the induction treatment was combined with radiation of the tumour area.. 24 patients were in complete remission after treatment (12 of these had all the metastases surgically removed before treatment). Nine patients had a partial remission while three patients had progressive disease. Median observation time was 49 months. 15 of 31 patients had an observation time of more than five years, and seven of these patients were alive after ten years. Three patients with metastasis localized to the lower extremity died of other causes. In one case the intra-arterial catheterisation caused serious aortic damage. INTERPRETATION: It is possible to achieve long-term remission in patients with metastatic malignant melanoma localized to an extremity after intra-arterial chemotherapy combined with radiotherapy.

[Malignant melanoma--diagnosis and treatment]. / Malignt melanom--diagnostikk og behandling.

BACKGROUND: Malignant melanoma is one of the most common forms of cancer in young adults in Norway. Early diagnosis and treatment is of vital importance for the prognosis. MATERIAL AND METHODS: The article is based on articles obtained from Medline and PubMed and from the Norwegian guidelines for the treatment of malignant melanoma. RESULTS AND INTERPRETATION: The depth of invasion of the melanoma is the most important prognostic factor. The thicker the primary tumour, the less chance there is of patient survival. Patients with a localized tumour should be controlled regularly for up to five years. Primary tumours, local recurrences and metastases should be treated by surgery whenever possible. Radical lymph node dissection should be performed in patients with regional lymph node metastases with the intention of curing the patient. There is no curative chemotherapy for metastatic disease. Dacarbazine has been the most used medication for many years. The treatment is effective for about 10 - 20 % of the patients, but the response does not last for long and the treatment does not prolong survival time. Isolated limb perfusion with melphalan and TNF-alpha (tumour necrosis factor-alpha) gives high response rates in patients with metastatic disease localised in the extremities, but the treatment is only palliative. Radiation therapy has been useful in selected patients postoperatively (e.g. after removed local recurrence) and in palliation of metastatic disease. Several clinical studies with chemotherapy, immunotherapy, vaccines and combinations are ongoing. Of preventive measures, avoiding sunburn is of particular importance.

Gonadal function and fertility in patients with bilateral testicular germ cell malignancy.

OBJECTIVE: To evaluate gonadal function and fertility in patients with bilateral testicular cancer (TC). METHODS: In 1999, 63 patients with bilateral invasive TC or carcinoma in situ (CIS) in the contralateral testis completed a mailed questionnaire evaluating their fatherhood (Cases). Their gonadal function had also been assessed after the first orchiectomy for TC before further treatment. The results were compared with those from 174 patients with unilateral TC (Controls). RESULTS: In Cases the post-orchiectomy serum levels of FSH and LH were above those of the Controls (p<0.001). Serum testosterone was similar, whereas sperm concentrations were lower in Cases (p<0.001). In Cases with metachronous invasive TC the level of serum FSH was associated with the interval between the two diagnoses. After the first orchiectomy, 10 of 25 Cases (40%) initiated a pregnancy, in 4 Cases by assisted fertilization. In the Control group 74% of the patients who attempted fatherhood succeeded (p=0.002). CONCLUSIONS: After unilateral orchiectomy for TC elevated serum FSH and/or oligospermia represent a high-risk factor of metachronous bilateral TC or synchronous CIS. At least one-third of these patients attempting fatherhood are successful after the first orchiectomy. Assisted fertilization is often necessary and the overall paternity rate is below that of patients with unilateral TC.