United States

Background
Numerous studies have attempted to construct cell classifiers based on machine learning in recent years. The following project examines the feasibility of building deep learning models to accurately distinguish cells that are nearly morphologically indistinguishable. The dataset used in the project is the International Symposium on Biomedical Imaging (ISBI) 2019 C-NMC dataset. The image repository was released to the public as part of a global machine learning competition that culminated in a symposium where the best performing models were presented. The cells were collected through bone marrow biopsies from 69 patients suffering from ALL and 49 healthy patients . Lymphoblasts within this dataset were digitally isolated and displayed on a black background.

Methods
Three different neural network models were created: a simple Logistic Regression model, a Convolutional Neural Network (CNN), and a ResNet50V2 neural network. To yield better results, simple adjustments to standard neural network parameters were implemented. The original dataset was imbalanced, so weights were applied to balance the dataset. Transformations, such as mirroring and rotation, were incorporated to increase the size of the dataset, which was necessary to address the limited dataset size. 

Results
The results of these models were promising with the highest performing model, CNN, accomplishing an accuracy score of 75%.

Conclusion
Ultimately, this branch of research is worth exploring further. Tuning parameters during model construction directly resulted in accuracy improvements. This has been proven in other studies used on the same dataset, which were able to accomplish higher accuracy scores based on different tuning methodologies.

AMA Research Challenge 2022

Diagnosing Acute Lymphoblastic Leukemia with Machine Learning

Naval Information Warfare Center Pacific is partnering with AFCEA San Diego chapter to host the workshop.

The NAML workshop is an opportunity for researchers from Navy research and development centers and our partner organizations to exchange technical ideas and to come together as a team.

The goals of NAML are:

To spread awareness of current machine learning research relevant to Naval applications

To connect machine learning researchers with experts in military requirements

To build and strengthen collaborations within the defense research community


<figure class="video_container">
 <iframe src="https://s3.amazonaws.com/pf-user-files-01/u-59356/uploads/2022-03-22/hk13mxo/NAML_ReceptionWithtitle.mp4" width=826 height=526 frameborder="0" allowfullscreen="true"> </iframe>
</figure>

Please log in to explore this event. You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

Registration for this workshop has now closed. We appreciate your interest and hope to welcome you at a future event. If you’d like to hear about similar upcoming events, please check our website or join our mailing list.

Registration for this workshop has now closed.

NAML 2022

Naval Information Warfare Center Pacific is partnering with AFCEA San Diego chapter to host the workshop. The NAML workshop is an opportunity for researchers from Navy research and development centers and our partner organizations to exchange technical ideas and to come together as a team.

The goals of NAML are:

To spread awareness of current machine learning research relevant to Naval applications

To connect machine learning researchers with experts in military requirements

To build and strengthen collaborations within the defense research community



poster

Welcome to the largest national, multi-specialty research event **featuring research from 750 medical students, residents and fellows, and international medical graduates.**


#### Event Details
The 2022 American Medical Association Research Challenge includes a poster symposium hall and a semifinals hall with the top scored research. Log in at any time throughout the event to:
* **View posters and presentations** on a variety of topics and specialties
* **Score posters in the semifinals** to help decide the top five finalists (AMA member exclusive opportunity)
* **Connect with participants** and provide feedback to help advance their research

At the conclusion of this event, the AMA will announce the top five finalists who will advance to the AMA Research Challenge finals, present their research to an elite panel of judges and compete to win a $10,000 grand prize sponsored by Laurel Road.
 

#### Attend Live Forums
**Oct. 20, 6 — 7 p.m. Central**
 **How to plan a career in academic medicine** 
 Learn how to plan a career in academic medicine and get your questions answered. [Access Here](https://underline.io/events/363/lounge)
 
 **Speakers**
![](https://assets.underline.io/uploads/markdown_image/1/image/d5adb133dfc58eb0a6d469e9cedcd5d4.jpg)
Ricardo Correa, MD, EdD
![](https://assets.underline.io/uploads/markdown_image/1/image/5f9437ab87861538bf02b9ca7c0d926e.jpg)
Stuart M. Greenstein, MD
![](https://assets.underline.io/uploads/markdown_image/1/image/05b1b637c99286d558c371f8e8b185ce.jpg)
Stephanie Strohbeen, MD
 
 **Oct. 22, 11 a.m. — Noon Central**
 **How to get published**
 Learn tips to get your research published and get your questions answered. [Access Here](https://underline.io/events/363/lounge)
 
 **Speakers**
![](https://assets.underline.io/uploads/markdown_image/1/image/8ce6235e957d31cb5e6f9097cc6f6972.jpg)
Jessica Li, MD/PhD Candidate
![](https://assets.underline.io/uploads/markdown_image/1/image/9950ef2e0237f54cb3d8280334d42491.jpg)
Charles Lopresto, DO

### Sponsor
AMA would like to thank Laurel Road for sponsoring a $10,000 grand prize for the winner of the AMA Research Challenge. Laurel Road is not involved in the judging or scoring process to select the winner. The winner will be announced on Dec. 7 during the AMA Research Challenge finals. Payment of the grand prize will be directly from Laurel Road. See the AMA Research Challenge Official Rules for terms and conditions.

![](https://assets.underline.io/uploads/markdown_image/1/image/611e331d4a0eba0920f240cfc04b4d77.png)

**There is no registration fee for this event! **

Register to explore research in a variety of topics. AMA members get exclusive access to score posters in the semifinals.

**Please use the link below to register for free if you have not already done so:**

https://www.ama-assn.org/about/events/ama-research-challenge

Registration Is Required

Welcome to the largest national, multi-specialty research event featuring research from 750 medical students, residents and fellows, and international medical graduates.



Introduction: Lung adenocarcinoma is the most prevalent non-small cell lung cancer. Up to 15% of lung adenocarcinomas are characterized by mutations in the epidermal growth factor receptor (EGFR) gene. Osimertinib, an EGFR tyrosine kinase inhibitor, is an effective method of treatment in EGFR-mutant lung cancer. However, osimertinib resistance is frequent. The mechanisms of Osimertinib are heterogenous but here we describe a new case of osimertinib resistant carcinosarcoma. 
Case history: A 63-year-old woman presented with headaches, right arm weakness, and gait instability. Radiographic evaluation showed a left parietal lobe mass, enlarged hilar lymph nodes, and numerous lung nodules. CT-guided biopsy of one lung nodule revealed an invasive adenocarcinoma which expressed TTF-1 but was negative for p-40 immunostaining. Tumor next generation sequencing revealed a rare exon 19 EGFR (E747_A750del) deletion. The patient was treated with stereotactic radiosurgery for brain metastasis followed by osimertinib immunotherapy. More than a year later, lead to a recurrence of the patient’s symptoms. Repeat radiographic evaluation showed new mediastinal lymphadenopathy. Fine needle aspiration of the enlarged mediastinal lymph nodes revealed carcinosarcoma with minimal adenocarcinoma component. A repeat tumor next generation sequencing confirmed the same EGFR exon 19 deletion (E746_A750del) in addition to TP53/Rb mutation and a high tumor mutational burden. 
Conclusion: This case demonstrates another association of osmertinib resistance in pulmonary adenocarcinoma with a specific EGFR mutation. This case is unique as it also documents histologic transformation of EGFR-mutated lung adenocarcinoma into a carcinosarcoma as a possible underlying mechanism of osimertinib resistance. 


EGFR-mutated lung adenocarcinoma transformation to carcinosarcoma as an osimertinib resistance mechanism

Abstract Title: Characterizing Outcomes of Immune Checkpoint Inhibition Treatment in Patients with advanced Renal Cell Carcinoma

Michael Pannell1, Maishara Muquith1, David Hsieh2, MD
1. University of Texas Southwestern Medical School
2. University of Texas Southwestern, Internal Medicine, Division of Hematology-Oncology

Background: Immune checkpoint inhibitors (ICIs) have become a standard of care treatment for patients with metastatic renal cell carcinoma (mRCC), with previous retrospective studies demonstrating potential benefits with increased tumor sensitivity to sequential treatment after ICI failure. An increasingly important area of investigation is characterizing the outcomes of ICI treatment and subsequent therapies in patients with advanced renal cell carcinoma. We examine characteristics of ICI treatment outcomes in a real-world cohort.

Methods: This IRB approved retrospective cohort study included patients with advanced renal cell carcinoma treated with palliative-intent PD-1/PD-L1 inhibitors at a single institution between 2016 and 2019. Clinical data including age, sex, cancer type, and treatment lines were obtained from the electronic medical record. Radiographic responses were determined from available CT/MRI studies and categorized using Response Evaluation Criteria In Solid Tumors 1.1 criteria. Response categories include complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD).

Results: Of 110 patients with mRCC that initiated ICI therapy between 2016 and 2019, 95 were treated with palliative intent and had subsequent imaging to evaluate their response. The overall objective response rate was 18.8% (n=18), with 18 patients demonstrating a partial response to treatment, 26 demonstrating stable disease, and 52 showing disease progression. ICI therapy as 1st line palliative treatment demonstrated the highest objective response rate of 27.5% (n=14), decreasing for subsequent treatment lines (13% and 0% for 2nd and 3rd+). Progression Free Survival (PFS) and Overall Survival (OS) were measured across ICI type and treatment line, with overall median values being 82 and 926 days, respectively. Additionally, patients were treated with either nivolumab monotherapy, nivolumab/ipilimumab combination, or pembrolizumab concurrent axitinib. Those treated with nivolumab monotherapy had a median PFS=79.5 days (n=32) and a median OS=1068.5 days (15 cases ongoing at the time of data analysis), compared to nivolumab/ipilimumab median PFS=83 days (n=55) and median OS=706 days (22 ongoing), and pembrolizumab median PFS=169 days (n=9) and median OS=559 (4 ongoing).

Conclusion: This study suggests that specific ICI therapies and treatment line may be associated with objective response and survival in patients with mRCC. Further identifying these associations could result in improved clinical outcomes.

Immune Checkpoint Inhibition Treatment Outcomes in Advanced Renal Cell Carcinoma

Background: Impact of race in advanced stage non-small cell lung cancer (NSCLC) patients treated with immune checkpoint inhibitors (ICIs) is conflicting. Few studies have significant representation of Black patients, and these studies lack detail on access to care, such as time to treatment initiation (TTI). Our study sought to examine racial disparities in TTI, overall survival (OS), and progression-free survival (PFS) using a population that was almost equally Black and White. 

Methods: This was an IRB approved retrospective cohort study of stage IV NSCLC patients > 18 years receiving immunotherapy at our center between 2014 and 2021. Associations between demographic characteristics and median values were examined using Chi-square tests. Kaplan-Meier curves and multivariate Cox proportional hazards model determined predictors of OS and PFS. Analyses were done using IBM PSAW (SPSS v.28) with p-value < 0.05 indicating significance.

Results: Of 194 patients who met the eligibility criteria, 42.3% were Black (n=82). On Kaplan-Meier analysis, no difference was observed between White and Black patients with respect to median PFS (6.30 versus 7.90 months; log-rank p=0.977) and OS (9.50 versus 11.77 months; log-rank p=0.457). On multivariate analysis, there was no difference in PFS (HR: 0.98; 95% CI: 0.68,1.43; p=0.928) or OS (HR: 1.01; 95% CI: 0.67,1.51; p=0.969) when controlling for race, sex, age at diagnosis, pack years, marital status, insurance type, histology, and Charlson Comorbidity Index (CCI). There was no difference in treatment selection between White and Black patients (p=0.184; chi-square), nor was there a difference in median time to overall treatment initiation (28.00 days versus 22.00 days; p=0.159). A difference was evident for immunotherapy initiation (34.00 versus 27.00 days; p=0.042) between White and Black patients respectively.

Conclusion: Although race was expected to permeate patient outcomes, OS and PFS did not differ between Black and White patients. The significantly shorter time to immunotherapy treatment for Black patients is an unexpected finding. In our study, we highlighted that equitable access reduces disparities, bridges gaps in care, and improves patient outcomes, regardless of race. Further studies with equitable representation of diverse racial groups are warranted to better understand how race impacts cancer outcomes.

Impact of race on outcomes of advanced stage non-small cell lung cancer in an equal access setting

MSIDx: Pan-cancer detection of microsatellite instability through targeted microsatellite sequencing enables immunotherapy

Abstract Title: Non-responder prediction in patients with advanced urothelial carcinoma using machine learning.

Background: Post FDA approval immunotherapy for cancer becomes widely used for locally advanced or metastatic urothelial carcinomas (1) with 16.5% of patients suffering Grade 3-5 side effects (2) from treatment. The current investigation wanted to classify patients that will not respond to immune checkpoint inhibitors (ICI) and potentially prevent side effects for these patients.

Methods: Pretreatment normalized ribonucleic acid Sequencing (RNASeq), clinical and other data for patients with advanced UC receiving PD-1 axis inhibitors were downloaded from gene expression omnibus (GEO) datasets project PRJNA735749 (3). This data also included responses to treatment data. Responders were categorized as patients with either complete response (CR), partial response (PR), or stable disease (SD) while non-responders comprised patients with progressive disease (PD). With a 70/30 training/test data split we developed four distinct algorithms were developed using a complete or targeted gene expression profile (high vs low expression with non-RNASeq data; total mutational burden (TMB), clinical data, and FGFR mutational status. The performance of each of the algorithms was assessed by comparing a matric of test set accuracy, sensitivity, specificity, positive predictive values (PPV), negative predictive values (NPV), and area under the receiver-operating curves.

Results: A total of 89 patients (from a total of 103) had complete RNASeq, DNA, and clinical data. The complete RNASeq + non-RNASeq data showed test set; accuracy (18.5%), sensitivity (100%), specificity (0.0%), PPV (25%), NPV (UD*). The targeted RNASeq + non-RNA Seq data showed test set; accuracy (70.4%), sensitivity (0.0%), specificity (100%), PPV (undefined), NPV (70.4%). The targeted RNASeq (Hi/Low cutoff) + non- RNASeq data using a Multilayer Perceptron (MLP) classifier showed test set; accuracy (81.5%), sensitivity (20.0%), specificity (95.5%), PPV (50.0%), NPV (84.0%). The targeted RNASeq (Hi/Low cutoff) + non-RNASeq data using a TensorFlow (TF) classifier showed test set; accuracy (77.8%), sensitivity (20.0%), specificity (90.9%), PPV (33.3%), NPV (83.3%).

Conclusion: The MLP/deep-learning classifier for targeted normalized high vs low RNA expression with TMB, clinical characteristics, and FGFR mutational status shows better overall results compared to other algorithms. However, these results need external validation and with a larger dataset and we may also be able to predict responders in these patients.


Non-responder prediction in patients with advanced urothelial carcinoma using machine learning

Background 
In patients diagnosed with cancer, pain can interfere with daily functioning, affecting overall quality of life. The predictors of pain in patients with cancer are not well-delineated. Further, pain has been associated with inflammation, yet previous research attempting to understand this link has been limited secondary to the lack of inclusion of other covariates associated with inflammation. The aims of this study were to describe pain intensity and interference in patients with cancer, investigate the predictors and consequences of pain, and examine the associations between pain and cytokines while covarying factors associated with inflammation. 

Methods
This study was a secondary analysis of baseline data from three prospective studies conducted between 2008-2019 at the University of Pittsburgh Medical Center Cancer Center. Patients completed questionnaires assessing sociodemographic characteristics, pain, depression, sleep, and fatigue. Serum levels of cytokines IL-1alpha, IL-1beta, TNF-alpha, IL-10, IL-2, and IFN-gamma were assessed. Analyses of Variance, Chi-square analyses, and linear regression were performed to test the hypotheses. 


Results
Of the 779 participants, the mean age was 63.5 years (S.D.=11.02), 57.8% male, 90.6% Caucasian. 67.3% of patients reported pain, of which 50.4% of patients reported was cancer-related. Significant predictors of pain intensity and interference included lower education level (beta=-2.800, 95% CI=[-4.374,-1.227], p=0.01) and a diagnosis of hepatocellular carcinoma or cholangiocarcinoma (beta=-0.947, 95% CI=[-1.551,-0.343], p=0.002). After adjusting for sociodemographic, disease-specific, and treatment-related factors, pain significantly predicted higher levels of depression (beta=0.587, 95% CI=[0.425,0.749], p<0.001), fatigue (beta=-0.858, 95% CI=[-1.031,-0.684], p<0.001), and fewer hours of sleep per night (beta=-0.055, 95% CI=[-0.088,-0.023], p=0.001). After adjusting for sociodemographic, disease-specific, and treatment-related factors, pain was not significantly related to serum levels of IL-1alpha, IL-1beta, TNF-alpha, IL-10, IL-2, or IFN-gamma.

Conclusion
Nearly one-third of patients reported cancer-related pain, which was linked to poorer quality of life. Pain, however, was not found to be associated with inflammatory biomarkers, warranting further research into underlying biological mechanisms of pain and related symptoms of depression, fatigue, and poor sleep. Given that pain predictors included lower education level and diagnosis of liver cancer, interventions should be designed and implemented for these patients with the greatest need for pain management.

Pain, Inflammation, and Psychosocial Factors in the Context of Cancer

Predictors of hematologic response and survival with stem cell transplantation in patients with AL amyloidosis: a 25-year longitudinal study

Background: RT recipients have a 5-6-fold increased risk of renal cell carcinoma (RCC). Post-RT RCCs usually arise from the native kidneys but may rarely develop in the allograft. We analyzed the incidence and clinical outcomes of patients with RA RCC at Memorial Hermann Hospital, Texas Medical Center.

Methods: We performed a single center retrospective study of patients with RA RCC who underwent transplantation between January 1, 1999, and December 31, 2019. The following data were collected: Baseline demographic variables, cause of end stage renal disease (ESRD), duration of dialysis prior to RT, interval from RT to RCC diagnosis, RCC stage, histologic subtype, and clinical outcome post RCC treatment.

Results: Among 1998 RT recipients, 7 patients (0.3%) developed RCC in the RA, of whom 4 were men. Subject races included White (n=3), Black (n=2), Hispanic (n=1), and Asian (n=1). The median age was 56 years (Range, 18-66). Causes of ESRD included diabetic or hypertensive nephropathy (n=3), congenital renal dysplasia (n=1), Alport syndrome (n=1), IgA nephropathy (n=1) and Henoch Schoenlein purpura (n=1). Median duration of dialysis prior to RT was 48 months (range, 1–120). Median time to RCC diagnosis post-RT was 108 months (range, 48-180). Histologies included clear cell (n=5) and papillary (n=2) RCC. Majority (n=6) were nonsmokers. 
Six patients (85%) had stage I disease and all of them had deterioration in renal function that prompted renal imaging and biopsy that led to RCC diagnosis. One patient had stage IV disease at diagnosis and presented with hematuria. All seven patients had evidence of chronic allograft nephropathy. One patient had BK Virus nephropathy in the allograft. Six patients underwent transplant nephrectomy while one had cryotherapy. Three patients went back on dialysis while three patients underwent a retransplantation. None had RCC recurrence post retransplantation. Three patients died: 1 each with sepsis, COVID 19 and metastatic RCC.

Conclusion: Although rare, RA RCC can adversely affect long term survival. Chronic inflammation in the setting of RA nephropathy and iatrogenic immunosuppression are believed to be the main driving factors. 



Renal Allograft (RA) Renal Cell Carcinoma in Renal Transplant (RT) Recipients: A Single Center Retrospective Review

Abstract Title:
Does definitive local therapy have a role in select HER2+de novo metastatic breast cancer patients treated with dual anti HER2 blockade?

Background:
The role of surgery with curative intent in HER2+de novo metastatic breast cancer (dnMBC) is uncertain in the era of dual antibody therapy. We sought to determine from existing retrospective data current practice patterns if an association exists between surgery to the primary tumor and improved survival in HER2+dnMBC patients treated with dual anti-HER2 blockade, accounting for selection bias.

Methods:
This study employed data from the National Cancer Database (NCDB) from the years 2013 to 2015. Study inclusion was limited to adult women with HER2+dnMBC, who received immunotherapy/biologic response modifier drugs (BRM) as a first line treatment. Patients who received both systemic therapy and surgery to the primary breast tumor and patients who received systemic therapy alone were analyzed in two groups. Chi-square test for discrete variables and Wilcox on RankSum test for numeric variables was used to compare the two groups based on patient, tumor, and treatment characteristics. The primary endpoint was overall survival from the time of diagnosis to the time of death.

Results:
928 women with HER2+dnMBC treated with BRM were identified with 43.5% (n=404) receiving surgery and 56.5% (n=524) receiving systemic therapy alone. The 3-year overall survival was superior for the surgery group (74.1%, 95% CI 67.9–79.2%) compared to the no surgery group (53.3%; 95% CI 47.6–58.6%). The no surgery group had median overall survival of 39.8 months (95% CI 34.1–44.9), while the surgery group had not yet reached median overall survival.

Conclusion:
In a group of HER2+dnMBC patients receiving systemic treatment in the era of dual antibody therapy, patients who underwent surgery had a superior 3-year survival rate than those who did not. There may be a role for a prospective trial in HER2+dnMBC patients with an excellent response to dual HER2 blockade to investigate the contribution of curative intent local therapy to the primary tumor compared to systemic therapy alone.


Should we operate?: The role of surgery in HER2+ de novo metastatic breast cancer treated with dual anti-HER2+ blockade

Background
Adolescent and Young Adult (AYA) cancer patients have been historically understudied. Many improvements in survival observed for pediatric and older adult patients have not been paralleled in the AYA population. Contributing factors of this discrepancy may involve unique challenges AYA patients face, including poor investigative research, limited access to care, or financial insecurity. Several studies have identified that these challenges disproportionally impact AYA patients belonging to racial minority and lower socioeconomic (SES) groups, thus leading to worsened survival outcomes. We aimed to investigate differences in survival among AYA patients of different races, ethnicities, and SES treated within the same health care system, Kaiser Permanente Southern California (KPSC).

Methods
We identified patients, aged 15-39, diagnosed with the fourteen most common cancers in this age group at KPSC between the years 2010-2018, yielding 6,463 patients. Cancers included were breast, testis, leukemia, Non-Hodgkin and Hodgkin lymphoma, melanoma, colon, cervical, uterine, renal, brain, soft tissue, rectosigmoid, ovarian, stomach, and bone. The Neighborhood Deprivation Index (NDI), a validated measure of socioeconomic deprivation obtained from geocoded data of home addresses, was used as a marker of SES. Race, ethnicity, and survival data were obtained from the Electronic Health Record. We examined differences in survival rates for all cancer types, and breast cancer and leukemia separately, for NDI quartiles and racial and ethnic groups. 

Results
Patients residing in the two most deprived areas have a higher all-cause mortality rate compared to those living in the two least deprived areas for all cancer types combined (p=0.0012). We did not detect a significant difference in survival rate between different NDI groups for patients with leukemia. White patients, compared to non-White patients, had statistically significant improved survival rates (p=0.0000036) for all cancer types combined. Black and Hispanic patients with leukemia had lower survival rates than Asian/Pacific Islander and White subjects with leukemia (p=0.0052). No difference was found among breast cancer patients of different NDI quartiles or races/ethnicities.

Conclusion
Patients experiencing higher levels of deprivation have lower survival rates compared to less deprived patients. White patients had significantly higher survival rates than Black, Hispanic, or Asian/Pacific Islander patients for all cancer types combined.


Premium content

Downloads

Next from AMA Research Challenge 2022

EGFR-mutated lung adenocarcinoma transformation to carcinosarcoma as an osimertinib resistance mechanism

Similar lecture

Non-critical Marine Application of Collaborative Learning with Spiking Neural Networks