Skip to content

non-hodgkin-lymphoma

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

non-hodgkin-lymphoma

README.md

Non-Hodgkin Lymphoma (DLBCL) — QSP Model

Disease: Diffuse Large B-Cell Lymphoma (DLBCL) | Category: Hematologic Oncology
Directory: non-hodgkin-lymphoma/ | Abbreviation: nhl
Date Added: 2026-06-25

DLBCL QSP Mechanistic Map

1. Disease Overview

Diffuse Large B-Cell Lymphoma (DLBCL) is the most common aggressive B-cell lymphoma, accounting for 25–30% of all non-Hodgkin lymphoma cases worldwide (~150,000 new cases/year globally). It arises from germinal center (GC) B-cells or post-GC activated B-cells and can be cured with R-CHOP in ~60–65% of patients. The remaining ~35–40% who relapse/progress have poor outcomes (median OS ~6 months).

Molecular Subtypes (Cell of Origin — COO)

Subtype Frequency Key Drivers R-CHOP Outcome
GCB-DLBCL ~50% BCL-6+, t(14;18)/BCL-2, EZH2 Y641 Better (~60% 5yr OS)
ABC-DLBCL ~35% MYD88 L265P, CD79A/B, NF-κB ↑ Worse (~40% 5yr OS)
Double-Hit ~8% MYC + BCL-2/BCL-6 rearrangement Very poor (~25% 5yr OS)
PMBL ~3% JAK-STAT, CD30+, 9p24.1 ampl Intermediate

2. Mechanistic Map

File: nhl_qsp_model.dot → nhl_qsp_model.svg / nhl_qsp_model.png

The mechanistic map contains 14 clusters and 120+ nodes covering:

Cluster Content
1. B-Cell Development Pro-B → Nadir B → GC B → Memory/Plasma; PAX5, BCL6, BLIMP1, IRF4, AID
2. GC Reaction & DLBCL Origin Dark/Light zone, FDC, Tfh, SHM, CSR; GCB/ABC/PMBL origin
3. BCR Signaling (PI3K/AKT/mTOR) Lyn→Syk→BTK→PLCγ2→Ca²⁺/NFAT; PI3K→AKT→mTORC1; RAS→MAPK
4. NF-κB (Canonical/Non-canonical) CBM complex→IKK→NF-κB p65/p50; NIK/BAFF-R non-canonical; MYD88 L265P
5. MYC & Cell Cycle c-MYC/t(8;14), Cyclin D/CDK4/6, RB1, E2F, p53/MDM2, DHL
6. BCL-2 Family & Apoptosis BCL-2/XL/MCL-1 vs BAX/BAK/BIM; Cytochrome C→Apaf1→Caspase cascade
7. Epigenetics EZH2 Y641, CREBBP/EP300, KMT2D, SWI/SNF, HDAC
8. Tumor Microenvironment TAM M1/M2, Tfh, NK, CD8+ CTL, Treg, CAF, VEGF, HIF-1α
9. Immune Evasion PD-1/PD-L1 (9p24.1), CD47/SIRPα, MHC-I loss (β2M), CD19/CD20 loss
10. Rituximab PK/PD 2-cmt + TMDD; CD20 binding; ADCC/CDC/direct apoptosis
11. CHOP Components Cyclophosphamide (4-OH-CPP), Doxorubicin, Vincristine, Prednisone
12. Novel Agents Polatuzumab vedotin (ADC), Ibrutinib (BTK), Venetoclax (BCL-2), Tafasitamab, Glofitamab
13. CAR-T Therapy Axi-cel/Tisa-cel/Liso-cel; CAR expansion; CRS; ICANS; CD19 loss resistance
14. Clinical Endpoints Lugano CR/PR/SD/PD; IPI/R-IPI; ctDNA MRD; PFS, OS, EFS24

3. mrgsolve ODE Model

File: nhl_mrgsolve_model.R

Model Structure (22 Compartments)

PK — 10 Compartments:

Compartment Description
RuxCent Rituximab central (mg)
RuxPeriph Rituximab peripheral (mg)
CD20_free Free CD20 antigen on tumor surface
CD20_RTX CD20-Rituximab complex (TMDD)
CPP_active 4-OH-Cyclophosphamide active metabolite (mg)
DoxCent Doxorubicin central (mg)
DoxPeriph Doxorubicin peripheral (mg)
VEN_gut Venetoclax gut absorption (mg)
VEN_cent Venetoclax central (mg)
IBR_cent Ibrutinib central (mg)

PD — 12 Compartments:

Compartment Description
Tumor Tumor burden (logistic growth, T0=100 au)
BCR_signal BCR/NF-κB signal activation (0-1)
BCL2_occ BCL-2 occupancy by Venetoclax (0-1)
NK_cells NK cell pool (ADCC effectors)
CD8_cells CD8+ T-cell pool
ANC Absolute Neutrophil Count (×10⁹/L)
Resistance Drug resistance fraction (0-1)
Cum_RTX_dose Cumulative rituximab exposure
Tumor_resp Tumor response tracker
CRS_risk CRS risk index

Key Parameters

Parameter Value Source
Rituximab CL 0.008 L/h Berinstein 1998
Rituximab Vc 3.1 L Maloney 1994
RTX T½ (terminal) ~21 days Avivi 2014
Tumor growth rate (kg) 0.012/day (T2 ~58d) Calibrated
R-CHOP CR rate (Day 126) ~65% Coiffier 2002
kon(CD20-RTX) 0.27/nM/h Gibiansky 2014

Treatment Scenarios (6)

# Scenario Treatment Clinical Reference
1 Untreated DLBCL No treatment Natural history
2 R-CHOP ×6 (Standard) Rituximab + CHOP ×6 cycles Coiffier 2002 NEJM
3 Pola-R-CHP ×6 (POLARIX) Polatuzumab vedotin + R-CHP Tilly 2022 NEJM
4 R-CHOP + Ibrutinib (ABC) Ibrutinib + R-CHOP (ABC-DLBCL) Younes 2019 NEJM
5 R-CHOP + Venetoclax Venetoclax + R-CHOP (BCL-2 high) Morschhauser 2021 JCO
6 R-CHOP (Double-Hit) R-CHOP in DHL (high-risk subgroup) Dunleavy 2013 JCO

4. Shiny Dashboard

File: nhl_shiny_app.R

6-Tab Interactive Dashboard

Tab Content
1. Patient Profile Disease subtype (GCB/ABC/DHL), IPI score, genomic drivers, BCL-2 status, prognosis by subtype
2. Drug PK Rituximab 2-cmt + TMDD plot, CHOP concentration-time, CD20 receptor occupancy, add-on agent PK (venetoclax/ibrutinib), PK parameter table
3. Tumor Dynamics SPD time course, response classification (CR/PR/SD/PD), NK/CD8+ immune effector dynamics, BCR/NF-κB signal suppression
4. Clinical Endpoints Tumor reduction kinetics, endpoint summary table, waterfall plot (N=50), Lugano criteria reference, KM-like PFS curve
5. Scenario Comparison 6-arm tumor burden comparison, scenario parameters, Day-126 ORR stacked bar, ANC myelosuppression by arm, resistance development
6. Biomarkers & Toxicity ANC time profile, BCL-2 occupancy (venetoclax), CRS risk index, cumulative doxorubicin cardiotoxicity monitor, biomarker table

Key Interactive Controls

  • DLBCL Subtype: GCB / ABC / Double-Hit (changes BCR signal, growth rate, prognosis)
  • BSA: adjusts absolute doses for RTX, CPP, Dox
  • R-CHOP cycles: 4–8 cycles
  • Dose sliders: RTX, cyclophosphamide, doxorubicin per m²
  • Add-on agents: toggle venetoclax (400–1200 mg) or ibrutinib (280–840 mg)
  • Simulation horizon: 180–730 days

5. Model Calibration

Clinical Trial Treatment Primary Endpoint Observed Model Target
Coiffier 2002 NEJM R-CHOP vs CHOP 3-yr EFS 59% vs 38% CR Day 126 ~65%
Tilly 2022 NEJM Pola-R-CHP vs R-CHOP 2-yr PFS 76.7% vs 70.2% ~6.5% PFS advantage
Younes 2019 NEJM (PHOENIX) Ibr+R-CHOP EFS (ABC) HR 0.934 (ns) Modest ABG benefit
Morschhauser 2021 JCO (CAVALLI) VEN+R-CHOP ORR (BCL-2+) 88% vs 67% ~21% ORR gain
Neelapu 2017 NEJM (ZUMA-1) Axi-cel ORR (refractory) 82% CAR-T module

6. Running the Model

# Requirements
install.packages(c("mrgsolve","dplyr","ggplot2","tidyr","shiny",
                   "shinydashboard","DT","scales"))

# Run simulation (deterministic)
source("nhl_mrgsolve_model.R")

# Launch Shiny dashboard
shiny::runApp("nhl_shiny_app.R")

# Render mechanistic map (Graphviz required)
# dot -Tsvg nhl_qsp_model.dot -o nhl_qsp_model.svg
# dot -Tpng -Gdpi=150 nhl_qsp_model.dot -o nhl_qsp_model.png

7. References Summary

50 PubMed-linked references organized by:

  • Pivotal Clinical Trials (R-CHOP, POLARIX, PHOENIX, CAVALLI, ZUMA-1, JULIET)
  • Disease Biology & Molecular Subtypes
  • BCR/NF-κB Signaling
  • BCL-2 & Apoptosis
  • MYC & Double-Hit Lymphoma
  • Novel Agents & CAR-T
  • Rituximab/CHOP/Venetoclax/Ibrutinib PK-PD
  • Response Assessment (Lugano criteria, ctDNA MRD)
  • Epigenetics & Resistance

See nhl_references.md for full list.