The challenges facing deep learning-based catheter localization for ultrasound guided high-dose-rate prostate brachytherapy.

BACKGROUND: Automated catheter localization for ultrasound guided high-dose-rate prostate brachytherapy faces challenges relating to imaging noise and artifacts. To date, catheter reconstruction during the clinical procedure is performed manually. Deep learning has been successfully applied to a wide variety of complex tasks and has the potential to tackle the unique challenges associated with multiple catheter localization on ultrasound. Such a task is well suited for automation, with the potential to improve productivity and reliability. PURPOSE: We developed a deep learning model for automated catheter reconstruction and investigated potential factors influencing model performance. The model was designed to integrate into a clinical workflow, with a proposed reconstruction confidence metric to aid in planner verification. METHODS: Datasets from 242 patients treated from 2016 to 2020 were collected retrospectively. The anonymized dataset comprises 31,000 transverse images reconstructed from 3D sagittal ultrasound acquisitions and 3500 implanted catheters manually localized by the planner. Each catheter was retrospectively ranked based on the severity of imaging artifacts affecting reconstruction difficulty. The U-NET deep learning architecture was trained to localize implanted catheters on transverse images. A fivefold cross-validation method was used, allowing for evaluation over the entire dataset. The postprocessing software combined the predictions with patient-specific implant information to reconstructed catheters in 3D space, uniquely matched to the implanted grid positions. A reconstruction confidence metric was calculated based on the number and probability of localized predictions per catheter. For each patient, deep learning prediction and postprocessing reconstruction were completed in under 2 min on a nonperformance PC. RESULTS: Overall, 80% of catheter reconstructions were accurate, within 2 mm along 90% of the length. The catheter tip was often not detected and required extrapolation during reconstruction. The reconstruction accuracy was 89% for the easiest catheter ranking and decreased to 13% for the highest difficulty ranking, when the aid of live ultrasound would have been recommended. Even when limited to the easiest ranked catheters, the reconstruction accuracy decreased at distal grid positions, down to 50%. Individual implantation style was found to influence the frequency of severe artifacts, slightly impacting the model accuracy. A reconstruction confidence metric identified the difficult catheters, removed the observed individual variation, and increased the overall accuracy to 91% while excluding 27% of the reconstructions. CONCLUSIONS: The deep learning model localized implanted catheters over a large clinical dataset, with overall promising results. The model faced challenges due to ultrasound artifacts and image degradation distal to the probe, underlining the continued importance of maintaining image quality and minimizing artifacts. A potential workflow for integration into the clinical procedure was demonstrated, including the use of a confidence metric to predict low accuracy reconstructions. Comparison between models evaluated on different datasets should also consider underlying differences, such as the frequency and severity of imaging artifacts.

Frequency of Infectious Mortality at the End of Induction Chemotherapy in Acute Lymphoblastic Leukemia and Lymphoma Patients: Findings From a Tertiary Care Cancer Center.

Background and objective In low- and low-to-middle-income countries (LMICs), the incidence of treatment-related mortality (TRM) in patients with acute lymphoblastic leukemia (ALL) and lymphoblastic lymphoma (LBL) is up to 52%. This study aimed to determine the mortality rate at the end of the induction phase of the treatment among patients with ALL and lymphoma at a tertiary care cancer center. Methods This retrospective study analyzed outcomes after induction chemotherapy in pediatric patients with acute leukemia and lymphoma at a tertiary care cancer center from January 2015 to December 2016. Information regarding demographics, clinical characteristics, and laboratory investigations were extracted and reviewed. Results Of the total 160 patients, 110 were males, and the mean age of the sample was 4.6 +2.8 years. B-cell leukemia (pre-B-ALL) was diagnosed in 84% (n=134), while 10% (n=6) had acute T-cell leukemia (pre-T-ALL) and 6% (n=10) had lymphoma. Sixteen patients (10%) died within the defined induction period, with 14 deaths occurring due to infections and two deaths resulting from chemotherapy-related toxicity. Conclusion Based on our findings, there is a significant prospect of mortality from infections during induction chemotherapy in patients with pediatric hematological malignancies.

A Single Institution's Experience with Cytogenetic and MRD Outcomes in Pediatric Acute Lymphoblastic Leukemia.

OBJECTIVE: To determine the frequency of cytogenetic type and its significance in the prognostic outcome of the pediatric patients in acute lymphoblastic leukemia (ALL), aged 1 to 15 years, and also determine the importance of minimal residual disease (MRD) in the management of the condition. STUDY DESIGN: An observational study. PLACE AND DURATION OF STUDY: Pediatric Oncology Ward, Shaukat Khanum Cancer Hospital, Lahore, from January 2015 to July 2017. METHODOLOGY: Patients aged 1-15 years, diagnosed with ALL, were included. Studied variables were cytogenetic type and MRD outcome in patients with ALL. Patients under one year of age and more than 15 years, or those having comorbidities, were excluded. RESULTS: Total 150 patients' data were retrieved from the Hospital database. One hundred and thirty-three belonged to age 1 to 5 years group (89%) and 17 (11%) were in 5 to 10 years group. The mean age of the patient was 4.3 +3.1 years. One hundred and two (68%) were males; whereas, 48 (32%) were females. Pre B acute lymphoblastic leukemia was diagnosed in 139 (93%) patients and 11(7%) were diagnosed with Pre T acute lymphoblastic leukemia. Standard risk was observed in 120 (80%) patients and 30 (20%) patients were on high risk as per National Cancer Institute (NCI) Guidelines. Regimen A was used in 125 (83.3%), Regimen B in 16 (10.7%), and Regimen C in 9 (6%) patients. BCR-ABL was positive in 2 (1.30%), TEL-AML in 68 (45%), MLL in 5 (3.30%), and normal in 54 (36%). MRD at day 29 was negative in 40 (93%) and positive in 3 (7%). The karyotyping was done in 128 (85%) patients, out of which 68 (53%) were hyperploids, 41 (32%) euploid, and 19 (15%) were hypoploid. Death was observed in 22 (15%) patients. Nineteen (86%) deaths were due to fungal and bacterial sepsis; and disease-related deaths were noted in 3 (14%) patients. CONCLUSION: The role of MRD and cytogenetics in risk assessment has improved in the early prognosis determination.

Fungal Infections In Paediatric Patients With Acute Lymphoblastic Leukaemia While On Induction Chemotherapy.

BACKGROUND: Acute Lymphoblastic Leukaemia (ALL) is one of the most common haematological malignancies seen in children. Despite steadily improving long-term outcomes, infections remain a major cause of morbidity and mortality in children receiving therapy for leukaemia. The incidence and risk of invasive fungal infections (IFIs) continue to rise. In some settings, IFIs caused by moulds are more frequent than those caused by yeasts, and Aspergillus spp. is the most common pathogens. The aim of the study was to determine the frequency, type of fungal infection seen during induction chemotherapy and outcomes.. METHODS: This observational retrospective study was conducted in paediatric oncology department at Shaukat Khanum Cancer Hospital Lahore from January 2015 to December 2016 after taking International research board (IRB) approval. The study includes all the patients aged 1-15 who were diagnosed with acute lymphoblastic leukaemia while on induction chemotherapy. The data was retrieved of 165 patients from the hospital database after informed written consent.. RESULTS: The mean age of the patients was 4.6±2.80 with range 1-15years. Total 154 (93%) of the children were of age between 1-5 years whereas only 11 (6.7%) were between 5-10 years. Male sex was predominant in 117 (70.9%) and 48 (29.1%) were girls. Pre-B Acute lymphoblastic leukaemia was diagnosed in 93.3% of the patients and rest 11 (6.7%) were diagnosed with Pre-T Acute lymphoblastic leukaemia. NCI Standard risk patients were 132 (80%) and 33 (20%) were stratified as high risk. Fungal infections were documented in 18 (11%) patients out of which 7(39%) were probable infections, and only 11 (61%) were proven fungal infections. Aspergillus was the commonest organism in 5 (28%) patients. Death was observed in 21 (13%) patients and causes were sepsis due to infections in 18 (86%) out of which fungal infections were 11(61%), bacterial 4(22%), combine bacterial and fungal 3(17%). Remaining 3 (14%) patients death causes include, neutropenic colitis was observed in first patient, second patient died of infection without any identifiable focus, and third patient died due to chemotherapy related toxicity.. CONCLUSIONS: Our study concludes that the fungal infection was the most common cause of mortality in induction in our patients. A prospective study in the form of clinical trial is needed to see if use of prophylactic antifungal can improve outcomes in our setting.

Luteinizing hormone-releasing hormone antagonists for urinary obstruction in prostate cancer.

Luteinizing hormone-releasing hormone (LHRH) antagonists rapidly reduce testosterone and are preferred to LHRH agonists in situations when early response is important. The lack of flare reaction, as compared to LHRH agonists, is particularly desirable as it would not aggravate the problem. A 78-year-old man presented with symptoms of urinary tract obstruction. He had a prostate-specific antigen (PSA) of 91.3 ug/L and serum creatinine 146 umol/L. He had a large pelvic mass due to histologically confirmed prostate cancer, resulting in moderate left hydronephrosis and deteriorating renal function (serum creatinine of 163 umol/L). He was started on combined degarelix and bicalutamide on the day of consultation (day 0). The hydronephrosis resolved on the repeat computerized tomography scan performed on day 10. Serum creatinine normalized to under 130 umol/L on day 18. The PSA fell to 11 ug/L on day 18, 2.8 ug/L on day 28, and 0.5 ug/L on day 53. Therefore, LHRH antagonists are particularly useful in urgent situations. It is the preferred choice in these circumstances.