The Pancreatic Cancer Early Detection (PRECEDE) Study is a Global Effort to Drive Early Detection: Baseline Imaging Findings in High-Risk Individuals.

BACKGROUND: Pancreatic adenocarcinoma (PC) is a highly lethal malignancy with a survival rate of only 12%. Surveillance is recommended for high-risk individuals (HRIs), but it is not widely adopted. To address this unmet clinical need and drive early diagnosis research, we established the Pancreatic Cancer Early Detection (PRECEDE) Consortium. METHODS: PRECEDE is a multi-institutional international collaboration that has undertaken an observational prospective cohort study. Individuals (aged 18-90 years) are enrolled into 1 of 7 cohorts based on family history and pathogenic germline variant (PGV) status. From April 1, 2020, to November 21, 2022, a total of 3,402 participants were enrolled in 1 of 7 study cohorts, with 1,759 (51.7%) meeting criteria for the highest-risk cohort (Cohort 1). Cohort 1 HRIs underwent germline testing and pancreas imaging by MRI/MR-cholangiopancreatography or endoscopic ultrasound. RESULTS: A total of 1,400 participants in Cohort 1 (79.6%) had completed baseline imaging and were subclassified into 3 groups based on familial PC (FPC; n=670), a PGV and FPC (PGV+/FPC+; n=115), and a PGV with a pedigree that does not meet FPC criteria (PGV+/FPC-; n=615). One HRI was diagnosed with stage IIB PC on study entry, and 35.1% of HRIs harbored pancreatic cysts. Increasing age (odds ratio, 1.05; P<.001) and FPC group assignment (odds ratio, 1.57; P<.001; relative to PGV+/FPC-) were independent predictors of harboring a pancreatic cyst. CONCLUSIONS: PRECEDE provides infrastructure support to increase access to clinical surveillance for HRIs worldwide, while aiming to drive early PC detection advancements through longitudinal standardized clinical data, imaging, and biospecimen captures. Increased cyst prevalence in HRIs with FPC suggests that FPC may infer distinct biological processes. To enable the development of PC surveillance approaches better tailored to risk category, we recommend adoption of subclassification of HRIs into FPC, PGV+/FPC+, and PGV+/FPC- risk groups by surveillance protocols.

Impact of pre-analytical variables on deep learning accuracy in histopathology.

AIMS: Machine learning (ML) binary classification in diagnostic histopathology is an area of intense investigation. Several assumptions, including training image quality/format and the number of training images required, appear to be similar in many studies irrespective of the paucity of supporting evidence. We empirically compared training image file type, training set size, and two common convolutional neural networks (CNNs) using transfer learning (ResNet50 and SqueezeNet). METHODS AND RESULTS: Thirty haematoxylin and eosin (H&E)-stained slides with carcinoma or normal tissue from three tissue types (breast, colon, and prostate) were photographed, generating 3000 partially overlapping images (1000 per tissue type). These lossless Portable Networks Graphics (PNGs) images were converted to lossy Joint Photographic Experts Group (JPG) images. Tissue type-specific binary classification ML models were developed by the use of all PNG or JPG images, and repeated with a subset of 500, 200, 100, 50, 30 and 10 images. Eleven models were generated for each tissue type, at each quantity of training images, for each file type, and for each CNN, resulting in 924 models. Internal accuracies and generalisation accuracies were compared. There was no meaningful significant difference in accuracies between PNG and JPG models. Models trained with more images did not invariably perform better. ResNet50 typically outperformed SqueezeNet. Models were generalisable within a tissue type but not across tissue types. CONCLUSIONS: Lossy JPG images were not inferior to lossless PNG images in our models. Large numbers of unique H&E-stained slides were not required for training optimal ML models. This reinforces the need for an evidence-based approach to best practices for histopathological ML.

Patient-reported Quality of Life and Satisfaction With Cosmetic Outcomes After Breast Conservation and Mastectomy With and Without Reconstruction: Results of a Survey of Breast Cancer Survivors.

OBJECTIVE: Although breast conservation is therapeutically equivalent to mastectomy for most patients with early-stage breast cancer, an increasing number of patients are pursuing mastectomy, which may be followed by breast reconstruction. We sought to evaluate long-term quality of life and cosmetic outcomes after different locoregional management approaches, as perceived by patients themselves. METHODS: We surveyed women with a diagnosis of nonmetastatic breast cancer from 2005 to 2007, as reported to the Los Angeles and Detroit population-based Surveillance, Epidemiology, and End Results registries. We received responses from 2290 women approximately 9 months after diagnosis (73% response rate) and from 1536 of these 4 years later. We evaluated quality of life and patterns and correlates of satisfaction with cosmetic outcomes overall and, more specifically, within the subgroup undergoing mastectomy with reconstruction, using multivariable linear regression. RESULTS: Of the 1450 patients who responded to both surveys and experienced no recurrence, 963 underwent breast-conserving surgery, 263 mastectomy without reconstruction, and 222 mastectomy with reconstruction. Cosmetic satisfaction was similar between those receiving breast conservation therapy and those receiving mastectomy with reconstruction. Among patients receiving mastectomy with reconstruction, reconstruction type and radiation receipt were associated with satisfaction (P < 0.001), with an adjusted scaled satisfaction score of 4.7 for patients receiving autologous reconstruction without radiation, 4.4 for patients receiving autologous reconstruction and radiation therapy, 4.1 for patients receiving implant reconstruction without radiation therapy, and 2.8 for patients receiving implant reconstruction and radiation therapy. CONCLUSIONS: Patient-reported cosmetic satisfaction was similar after breast conservation and after mastectomy with reconstruction. In patients undergoing postmastectomy radiation, the use of autologous reconstruction may mitigate the deleterious impact of radiation on cosmetic outcomes.

Impact of adjuvant chemotherapy on long-term employment of survivors of early-stage breast cancer.

BACKGROUND: Many women with early-stage breast cancer are working at the time of diagnosis and survive without disease recurrence. The short-term impact of chemotherapy receipt on employment has been demonstrated, but the long-term impact merits further research. METHODS: The authors conducted a longitudinal multicenter cohort study of women diagnosed with nonmetastatic breast cancer between 2005 and 2007, as reported to the population-based Los Angeles and Detroit Surveillance, Epidemiology, and End Results program registries. Of 3133 individuals who were sent surveys, 2290 (73%) completed a baseline survey soon after diagnosis and of these, 1536 (67%) completed a 4-year follow-up questionnaire. RESULTS: Of the 1026 patients aged < 65 years at the time of diagnosis whose breast cancer did not recur and who responded to both surveys, 746 (76%) worked for pay before diagnosis. Of these, 236 (30%) were no longer working at the time of the follow-up survey. Women who received chemotherapy as part of their initial treatment were less likely to be working at the time of the follow-up survey (38% vs 27%; P = .003). Chemotherapy receipt at the time of diagnosis (odds ratio, 1.4; P = .04) was found to be independently associated with unemployment during survivorship in a multivariable model. Many women who were not employed during the survivorship period wanted to work: 50% reported that it was important for them to work and 31% were actively seeking work. CONCLUSIONS: Unemployment among survivors of breast cancer 4 years after diagnosis is often undesired and appears to be related to the receipt of chemotherapy during initial treatment. These findings should be considered when patients decide whether to receive adjuvant chemotherapy, particularly when the expected benefit is low.

Providers of follow-up care in a population-based sample of breast cancer survivors.

To describe which providers provide breast cancer survivorship care, we conducted a longitudinal survey of nonmetastatic breast cancer patients identified by the SEER registries of Los Angeles and Detroit. Multinomial logistic regression examined the adjusted odds of surgeon compared with a medical oncologist follow-up or primary care provider compared with medical oncologist follow-up, adjusting for age, race/ethnicity, insurance, tumor stage, receipt of chemotherapy, endocrine therapy use, and visit to a medical oncologist at the time of diagnosis. Results were weighted to account for sample selection and nonresponse. 844 women had invasive disease and received chemotherapy or endocrine therapy. 65.2 % reported medical oncologists as their main care provider at 4 years, followed by PCP/other physicians (24.3 %) and surgeons (10.5 %). Black women were more likely to receive their follow-up care from surgeons (OR 2.47, 95 % CI 1.16-5.27) or PCP/other physicians (OR 2.62, 95 % CI 1.47-4.65) than medical oncologists. Latinas were more likely to report PCP/other physician follow-up than medical oncologists (OR 2.33, 95 % CI 1.15-4.73). Compared with privately insured women, Medicaid recipients were more likely to report PCP/other physician follow-up (OR 2.52, 95 % CI 1.24-5.15). Women taking endocrine therapy 4 years after diagnosis were less likely to report surgeons or PCP/other physicians as their primary provider of breast cancer follow-up care. Different survivorship care patterns emerge on race/ethnicity and insurance status. Interventions are needed to inform patients and providers on the recommended sources of breast cancer follow-up.

Cancer incidence among Arab Americans in California, Detroit, and New Jersey SEER registries.

OBJECTIVES: We calculated cancer incidence for Arab Americans in California; Detroit, Michigan; and New Jersey, and compared rates with non-Hispanic, non-Arab Whites (NHNAWs); Blacks; and Hispanics. METHODS: We conducted a study using population-based data. We linked new cancers diagnosed in 2000 from the Surveillance, Epidemiology, and End Results Program (SEER) to an Arab surname database. We used standard SEER definitions and methodology for calculating rates. Population estimates were extracted from the 2000 US Census. We calculated incidence and rate ratios. RESULTS: Arab American men and women had similar incidence rates across the 3 geographic regions, and the rates were comparable to NHNAWs. However, the thyroid cancer rate was elevated among Arab American women compared with NHNAWs, Hispanics, and Blacks. For all sites combined, for prostate and lung cancer, Arab American men had a lower incidence than Blacks and higher incidence than Hispanics in all 3 geographic regions. Arab American male bladder cancer incidence was higher than that in Hispanics and Blacks in these regions. CONCLUSIONS: Our results suggested that further research would benefit from the federal recognition of Arab Americans as a specified ethnicity to estimate and address the cancer burden in this growing segment of the population.

Cancer-related concerns among women with a new diagnosis of gynecological cancer: an exploration of age group differences.

OBJECTIVE: The study aimed to characterize cancer-related concerns among women with a new diagnosis of gynecological cancer from a developmental life stage perspective. The study compared the degree of cancer-related concern between young women (45 years or younger), middle age women (46-64 years), and older women (65 years or older). MATERIALS/METHODS: Data from women (N = 243) with a condition diagnosed as primary gynecological cancer who were participating in a randomized control trial were analyzed. Women completed a measure that assessed the degree of concern in 12 cancer-related domains (physical functioning, cancer treatment, emotional functioning, sexual functioning, disease progression/death, own well-being, partner well-being, relationship with spouse/partner, body image, relationship with others, employment, and finances). Multivariate comparisons were made between the 3 age groups on the cancer-related concerns. RESULTS: There were age group differences in overall cancer-related concern and specific cancer-related domains. Young women reported the greatest cancer-related concern (P < 0.001). They reported greater concern over emotional functioning (P < 0.001) and sexual functioning (P < 0.001) compared to the middle- and older-age groups. Older women reported less concern over the impact of cancer on finances (P = 007). There were no differences between age groups in concern over physical impairment, cancer treatment, disease progression/death, own well-being, partner well-being, relationship with spouse/partner, body image, and relationship with others. CONCLUSIONS: Age may play an important role in the impact of a gynecological cancer diagnosis in domains of functioning, specifically emotional functioning, sexual functioning, and finances. Other cancer-related areas may represent more universal degree of impact. Professionals may benefit from considering the impact of cancer from a developmental life stage perspective.

Adjuvant endocrine therapy initiation and persistence in a diverse sample of patients with breast cancer.

Adjuvant endocrine therapy for breast cancer reduces recurrence and improves survival rates. Many patients never start treatment or discontinue prematurely. A better understanding of factors associated with endocrine therapy initiation and persistence could inform practitioners how to support patients. We analyzed data from a longitudinal study of 2,268 women diagnosed with breast cancer and reported to the Metropolitan Detroit and Los Angeles SEER cancer registries in 2005-2007. Patients were surveyed approximately both 9 months and 4 years after diagnosis. At the 4-year mark, patients were asked if they had initiated endocrine therapy, terminated therapy, or were currently taking therapy (defined as persistence). Multivariable logistic regression models examined factors associated with initiation and persistence. Of the 743 patients eligible for endocrine therapy, 80 (10.8 %) never initiated therapy, 112 (15.1 %) started therapy but discontinued prematurely, and 551 (74.2 %) continued use at the second time point. Compared with whites, Latinas (OR 2.80, 95 % CI 1.08-7.23) and black women (OR 3.63, 95 % CI 1.22-10.78) were more likely to initiate therapy. Other factors associated with initiation included worry about recurrence (OR 3.54, 95 % CI 1.31-9.56) and inadequate information about side effects (OR 0.24, 95 % CI 0.10-0.55). Factors associated with persistence included two or more medications taken weekly (OR 4.19, 95 % CI 2.28-7.68) and increased age (OR 0.98, 95 % CI 0.95-0.99). Enhanced patient education about potential side effects and the effectiveness of adjuvant endocrine therapy in improving outcomes may improve initiation and persistence rates and optimize breast cancer survival.

Surgeon-level variation in patients' appraisals of their breast cancer treatment experiences.

BACKGROUND AND PURPOSE: While variation in breast cancer quality indicators has been studied, to date there have been no studies examining the degree of surgeon-level variation in patient-reported outcomes. The purpose of this study is to examine surgeon-level variation in patient appraisals of their breast cancer care experiences. METHODS: Survey responses and clinical data from breast cancer patients reported to Detroit and Los Angeles Surveillance, Epidemiology and End Results registries from 6/2005 to 2/2007 were merged with attending surgeon surveys (1,780 patients, 291 surgeons). Primary outcomes were patient reports of access to care, care coordination, and decision satisfaction. Random-effects models examined variation due to individual surgeons for these three outcomes. RESULTS: Mean values on each patient-reported outcome scale were high. The amount of variation attributable to individual surgeons in the unconditional models was low to modest: 5.4% for access to care, 3.3% for care coordination, and 7.5% for decision satisfaction. Few factors were independently associated with patient reports of better access to or coordination of care, but less-acculturated Latina patients had lower decision satisfaction. CONCLUSIONS: Patients reported generally positive experiences with their breast cancer treatment, though we found disparities in decision satisfaction. Individual surgeons did not substantively explain the variation in any of the patient-reported outcomes.

Social-cognitive processes associated with fear of recurrence among women newly diagnosed with gynecological cancers.

OBJECTIVE: This cross sectional study aimed to characterize fears of recurrence among women newly diagnosed with gynecologic cancer. The study also evaluated models predicting the impact of recurrence fears on psychological distress through social and cognitive variables. METHODS: Women (N=150) who participated in a randomized clinical trial comparing a coping and communication intervention to a supportive counseling intervention to usual care completed baseline surveys that were utilized for the study. The survey included the Concerns about Recurrence Scale (CARS), Beck Depression Inventory (BDI), Impact of Event Scale (IES), and measures of social (holding back from sharing concerns and negative responses from family and friends) and cognitive (positive reappraisal, efficacy appraisal, and self-esteem appraisal) variables. Medical data was obtained via medical chart review. RESULTS: Moderate-to-high levels of recurrence fears were reported by 47% of the women. Younger age (p<.01) and functional impairment (p<.01) correlated with greater recurrence fears. A social-cognitive model of fear of recurrence and psychological distress was supported. Mediation analyses indicated, that as a set, the social and cognitive variables mediated the association between fear of recurrence and both depression and cancer-specific distress. Holding back and self-esteem showed the strongest mediating effects. CONCLUSION: Fears of recurrence are prevalent among women newly diagnosed with gynecologic cancer. Social and cognitive factors play a role in women's adaptation to fears and impact overall psychological adjustment. These factors may be appropriate targets for intervention.

A Case Report of a Patient With Prostate Adenocarcinoma Metastatic to the Posterior Peritoneum Despite the Negative Preoperative Prostate-Specific Membrane Antigen (PSMA) Positron Emission Tomography/Computed Tomography (PET/CT) Scan.

Despite the role of prostate-specific antigen (PSA) screening and the multitude of therapies available, prostate cancer (PCa) remains a leading cause of cancer-related morbidity and mortality. For many patients diagnosed with PCa, clinical and radiographic staging are critical components for management decisions. PCa staging with the use of imaging modalities such as MRI and bone scintigraphy is recommended in patients with newly diagnosed intermediate or high-risk PCa and in patients with biochemical recurrence; it is also recommended for monitoring the patient's response to treatment for diagnosed PCa. Prostate-specific membrane antigen (PSMA) positron emission tomography/computed tomography (PET/CT), recently approved in 2021, is an imaging modality that has been shown to have a greater sensitivity, specificity, and negative likelihood ratio than conventional imaging modalities such as CT, bone scintigraphy, and MRI in prostate cancer staging. Despite the improvement in staging that PSMA-PET/CT can provide, our current report details a false-negative result in detecting a rare PCa metastasis to the peritoneum, which was found at the time of an attempted radical prostatectomy. Although the patient had a negative preoperative PSMA-PET/CT and was presumed to be non-metastatic, the prostatectomy was aborted because the patient was unexpectedly found to have peritoneal metastasis.

Learning fast and fine-grained detection of amyloid neuropathologies from coarse-grained expert labels.

Precise, scalable, and quantitative evaluation of whole slide images is crucial in neuropathology. We release a deep learning model for rapid object detection and precise information on the identification, locality, and counts of cored plaques and cerebral amyloid angiopathies (CAAs). We trained this object detector using a repurposed image-tile dataset without any human-drawn bounding boxes. We evaluated the detector on a new manually-annotated dataset of whole slide images (WSIs) from three institutions, four staining procedures, and four human experts. The detector matched the cohort of neuropathology experts, achieving 0.64 (model) vs. 0.64 (cohort) average precision (AP) for cored plaques and 0.75 vs. 0.51 AP for CAAs at a 0.5 IOU threshold. It provided count and locality predictions that correlated with gold-standard CERAD-like WSI scoring (p=0.07± 0.10). The openly-available model can quickly score WSIs in minutes without a GPU on a standard workstation.

Learning fast and fine-grained detection of amyloid neuropathologies from coarse-grained expert labels.

Precise, scalable, and quantitative evaluation of whole slide images is crucial in neuropathology. We release a deep learning model for rapid object detection and precise information on the identification, locality, and counts of cored plaques and cerebral amyloid angiopathy (CAA). We trained this object detector using a repurposed image-tile dataset without any human-drawn bounding boxes. We evaluated the detector on a new manually-annotated dataset of whole slide images (WSIs) from three institutions, four staining procedures, and four human experts. The detector matched the cohort of neuropathology experts, achieving 0.64 (model) vs. 0.64 (cohort) average precision (AP) for cored plaques and 0.75 vs. 0.51 AP for CAAs at a 0.5 IOU threshold. It provided count and locality predictions that approximately correlated with gold-standard human CERAD-like WSI scoring (p = 0.07 ± 0.10). The openly-available model can quickly score WSIs in minutes without a GPU on a standard workstation.

Underascertainment of radiotherapy receipt in Surveillance, Epidemiology, and End Results registry data.

BACKGROUND: Surveillance, Epidemiology, and End Results (SEER) registry data have been used to suggest underuse and disparities in receipt of radiotherapy. Prior studies have cautioned that SEER may underascertain radiotherapy but lacked adequate representation to assess whether underascertainment varies by geography or patient sociodemographic characteristics. The authors sought to determine rates and correlates of underascertainment of radiotherapy in recent SEER data. METHODS: The authors evaluated data from 2290 survey respondents with nonmetastatic breast cancer, aged 20 to 79 years, diagnosed from June of 2005 to February 2007 in Detroit and Los Angeles and reported to SEER registries (73% response rate). Survey responses regarding treatment and sociodemographic factors were merged with SEER data. The authors compared radiotherapy receipt as reported by patients versus SEER records. The authors then assessed correlates of radiotherapy underascertainment in SEER. RESULTS: Of 1292 patients who reported receiving radiotherapy, 273 were coded as not receiving radiotherapy in SEER (underascertained). Underascertainment was more common in Los Angeles than in Detroit (32.0% vs 11.25%, P < .001). On multivariate analysis, radiotherapy underascertainment was significantly associated in each registry (Los Angeles, Detroit) with stage (P = .008, P = .026), income (P < .001, P = .050), mastectomy receipt (P < .001, P < .001), chemotherapy receipt (P < .001, P = .045), and diagnosis at a hospital that was not accredited by the American College of Surgeons (P < .001, P < .001). In Los Angeles, additional significant variables included younger age (P < .001), nonprivate insurance (P < .001), and delayed receipt of radiotherapy (P < .001). CONCLUSIONS: SEER registry data as currently collected may not be an appropriate source for documentation of rates of radiotherapy receipt or investigation of geographic variation in the radiation treatment of breast cancer.

Leukoerythroblastosis in a Sickle Cell Patient With Pregnancy: An Interesting Peripheral Blood Smear Finding.

The global pandemic of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has shaken the entire world. The social, health and financial impacts of this pandemic are beyond words. We have learnt a lot about this new disease in a short period of time, but still a long road to go to fully determine its pathogenic effect. The primary target of this virus is angiotensin-converting enzyme 2 (ACE2) receptor, which is prevalent in endothelial cells throughout the body. Immunocompromised patients such as patients with sickle cell disease are more vulnerable to severe respiratory infections, including infection with SARS-CoV-2. In addition, sickle cell disease patients are prone to vaso-occlusive crisis, and theoretically SARS-CoV-2 can worsen the situation as it also can cause endothelial dysfunction and thrombosis. Herein, we are sharing an interesting peripheral blood smear finding of an asymptomatic 31-year-old multigravida pregnant female with a history of sickle cell disease and found to have a positive COVID-19 polymerase chain reaction (PCR) test during her third trimester of pregnancy at a routine clinic visit. Two weeks after the initial positive test, she developed nausea, vomiting, constipation and a pain crisis affecting her extremities while her COVID-19 PCR test was still positive. She was hemodynamically stable, and lab workup revealed chronic anemia, leukocytosis with neutrophilia and lymphopenia. Morphologic examination of the peripheral blood smear showed a marked leukoerythroblastosis: rare myeloblasts, sickle cells, markedly abundant nucleated red blood cells (RBCs), metamyelocytes, and many large and giant platelets were seen. In this context, her previous peripheral blood smears (prior to positive COVID-19 test) did not show leukoerythroblastosis. She was managed conservatively with hydration and pain control and delivered at 36 weeks via cesarean section due to pre-term labor and intrauterine growth retardation. The unusual finding of leukoerythroblastosis in a pregnant sickle cell disease patient with an asymptomatic COVID-19 infection indicates further studies to determine its effect on hematopoietic system and elucidate its clinical significance.

Receipt of delayed breast reconstruction after mastectomy: do women revisit the decision?

BACKGROUND: Postmastectomy breast reconstruction is an important component of breast cancer care, but few receive it at the time of the mastectomy. Virtually nothing is known about receipt of reconstruction after initial cancer therapy and why treatment might be delayed. MATERIALS AND METHODS: A 5-year follow-up survey was mailed to a population-based cohort of mastectomy-treated breast cancer patients who were initially surveyed at time of diagnosis in 2002 and reported to the Los Angeles and Detroit SEER registries (N = 645, response rate 60%). Outcomes were receipt of reconstruction (immediate [IR], delayed [DR], or none) and patient appraisal of their treatment decisions. RESULTS: About one-third (35.9%) had IR, 11.5% had DR, and 52.6% had no reconstruction. One-third delayed reconstruction because they focused more on other cancer interventions, and nearly half were concerned about surgical complications and interference with cancer surveillance. Two-thirds of those with no reconstruction said that the procedure was not important to them. A large proportion of all patients were satisfied with their reconstruction decision-making (89.4% IR, 78.4% DR, 80.4% no reconstruction, P = NS). However, only 59.3% of those with no reconstruction felt that they were adequately informed about their reconstructive options (vs 82.7% IR and 78.4% DR, P < .01). CONCLUSIONS: There was modest uptake of breast reconstruction after initial cancer treatment. Factors associated with delayed reconstruction were primarily related to uncertainty about the procedure, concern about cancer surveillance, and low priority. Those without reconstruction demonstrated significant informational needs, which should be addressed with future research efforts.

Intracranial myeloid sarcoma presentation in distant acute myeloid leukemia remission.

Intracranial myeloid sarcoma (IMS) is a rare central nervous system manifestation of hematopoietic neoplasms of myeloid origin. We report the first case of IMS treatment with an isocitrate dehydrogenase-2 (IDH-2) inhibitor, Enasidenib, following surgical resection, whole-brain radiation, and consolidation Etoposide/Cytarabine therapy. A 42-year-old female was diagnosed with IMS after a 10-year remission of her acute myeloid leukemia (AML). She underwent surgical debulking and had postoperative resolution of her visual symptoms. She received adjuvant radiation and medical management, and continues to show no evidence of recurrence or progression at 17 months postoperatively. This case is notable for an isolated IMS presentation in a patient with a very distant history of AML remission, and without evidence of concurrent bone marrow relapse. The goals of neurosurgical intervention should be symptomatic relief of mass effect and pathological diagnosis, due to the sensitivity of IMS to adjuvant radiation and medical management such as IDH-2 inhibitors.

Effects of Image Quantity and Image Source Variation on Machine Learning Histology Differential Diagnosis Models.

AIMS: Histology, the microscopic study of normal tissues, is a crucial element of most medical curricula. Learning tools focused on histology are very important to learners who seek diagnostic competency within this important diagnostic arena. Recent developments in machine learning (ML) suggest that certain ML tools may be able to benefit this histology learning platform. Here, we aim to explore how one such tool based on a convolutional neural network, can be used to build a generalizable multi-classification model capable of classifying microscopic images of human tissue samples with the ultimate goal of providing a differential diagnosis (a list of look-alikes) for each entity. METHODS: We obtained three institutional training datasets and one generalizability test dataset, each containing images of histologic tissues in 38 categories. Models were trained on data from single institutions, low quantity combinations of multiple institutions, and high quantity combinations of multiple institutions. Models were tested against withheld validation data, external institutional data, and generalizability test images obtained from Google image search. Performance was measured with macro and micro accuracy, sensitivity, specificity, and f1-score. RESULTS: In this study, we were able to show that such a model's generalizability is dependent on both the training data source variety and the total number of training images used. Models which were trained on 760 images from only a single institution performed well on withheld internal data but poorly on external data (lower generalizability). Increasing data source diversity improved generalizability, even when decreasing data quantity: models trained on 684 images, but from three sources improved generalization accuracy between 4.05% and 18.59%. Maintaining this diversity and increasing the quantity of training images to 2280 further improved generalization accuracy between 16.51% and 32.79%. CONCLUSIONS: This pilot study highlights the significance of data diversity within such studies. As expected, optimal models are those that incorporate both diversity and quantity into their platforms.s.