Projects / Mukherjee Lab

Mukherjee Lab

Brain Tumor Outcomes, Palliative Care, and Computational Segmentation

NCDB analysis and segmentation work told through the project figures, tables, and source files.

Kaplan-Meier survival curves comparing academic and non-academic facility groups
Kaplan-Meier curves became the clearest visual summary: facility type, survival time, and cohort filtering all converge in one figure.
01
My RoleResearch assistant focused on data analysis and reproducible workflows
02
Location / DatesMukherjee Lab, Johns Hopkins University
2024 – Present
03
OutputsTables, PRISMA diagram, KM curves, tumor-grade figures, model scripts
04
Skills LearnedR, survival analysis, survminer, readxl, Python, PyTorch, reproducible workflows

Evidence Trail

Who entered the study?

PRISMA diagram showing cohort selection
The PRISMA flow is the starting point: it shows exactly how the database population narrowed into the analyzable cohort.

How was disease grouped?

Tumor grade analysis figure
Tumor-grade grouping set the comparison structure, keeping downstream survival and palliative-care analyses clinically grounded.

What changed over time?

Kaplan-Meier survival curve figure
The survival curves turn the cleaned cohort into the main takeaway: outcomes can be compared without rereading the full analysis script.

Table Signals

Care groups looked different

CharacteristicCurativePain ManagementP
Sample Size64,057380
Mean Age65 +/- 1171 +/- 12<0.001
Female43.0%47.4%0.101
Survival Months9.361.84<0.001
View source table

Predictors pointed to access and acuity

VariableAdjusted ORP
Increasing Age1.04 [1.02-1.05]<0.001
Highest Income Quartile0.66 [0.44-0.99]0.045
Charlson-Deyo 21.83 [1.33-2.47]<0.001
Right Laterality0.55 [0.43-0.71]<0.001
View analysis PDF

Survival models added context

VariableHazard RatioP
Age1.03 [1.03-1.04]<0.001
Male vs Female1.06 [1.04-1.07]<0.001
Highest Income Quartile0.81 [0.78-0.83]<0.001
Hispanic Ethnicity0.78 [0.75-0.81]<0.001
View data dictionary