HongKong Nasopharyngeal

Slide1: 

For demonstration of interpreting a strong positive interaction see slides 25-30.

NPC-9902 Trial on Therapeutic Gain by Concurrent Chemotherapy and/or Accelerated Fractionation for Locally-Advanced Nasopharyngeal Cancer: 

NPC-9902 Trial on Therapeutic Gain by Concurrent Chemotherapy and/or Accelerated Fractionation for Locally-Advanced Nasopharyngeal Cancer Anne W. M. Lee for Hong Kong Nasopharyngeal Cancer Study Group

Estimation on Sample & Eligible Patients: 

Estimation on Sample & Eligible Patients Incidence in Hong Kong: CT-staged patients

NPC-9902: Eligibility Criteria: 

NPC-9902: Eligibility Criteria Non-keratinizing / Undifferentiated type T3-4 N0-1 M0 (UICC 5th) Staging Ix: CT / MR X-ray / CT thorax + metastatic work-up

Eligibility Criteria : 

Eligibility Criteria

Slide6: 

CF CF+C Stratification by center, N-stage, T-stage Randomization (equal proportion) AF AF+C CF = conventional fractionation; AF = accelerated fractionation; +C = plus chemotherapy

RT: Same for all Arms: 

RT: Same for all Arms

RT: Fractionation: 

RT: Fractionation Fractionation dose: 2 Gy/ Fr

Treatment in CRT Arm: 

Treatment in CRT Arm Chemotherapy (IGS regimen)   Concurrent Adjuvant P 80 + 5FU 1000 x 4d (q4 wk x 3 cycles)

Statistical Consideration: 

Statistical Consideration Target accrual = 464 patients Hypothesis Improve Failure-Free Survival (FFS) [Time to failure (any site)] from 45% to 55% (5-year) (alpha = 0.05, power = 90%)

Other End Points: 

Other End Points Progression-Free Survival (PFS) [Time to first failure or death (any cause)] Overall Survival

Participating Centers: 

Participating Centers

Accrual Rate: 

Accrual Rate July 1999 – April 2004: 189 patients

Slide14: 

Premature Termination due to Slow Accrual Recommendation by Data Monitoring Committee

Results: 

Results Period: Jul 1999 – Apr 2004 Total No. randomized: 189 pt. Regular follow-up: 97% Median follow-up: 2.9 (0.1 – 5.5) yr

Patient Characteristics: 

Patient Characteristics

Tumor Staging: 

Tumor Staging

Radiotherapy: 

Radiotherapy

Comparison of Prognostic Factors: 

Comparison of Prognostic Factors RT vs CRT All balanced CF vs AF All balanced Arm Balanced except AF Arm vs other Arms: Male: 90% vs 70% (p = 0.046) PS 2: 4% vs 0% Non-Chinese: 4% vs 0%

Major Violation of Protocol: 

Major Violation of Protocol

Compliance to Chemotherapy: 

Compliance to Chemotherapy AF+C 84% 73% Concurrent cycles: > 2 Adjuvant cycles: > 3 CF+C 92% 64%

Acute Toxicity (Grade >3): 

Acute Toxicity (Grade >3)

Chemotherapy Toxicity in CRT Arm: 

Chemotherapy Toxicity in CRT Arm

Overall Acute Toxicity (Grade >3): 

Overall Acute Toxicity (Grade >3)

Overall Acute Toxicity (Grade >3): 

Overall Acute Toxicity (Grade >3)

Failure-Free Survival (3-year): 

Failure-Free Survival (3-year) RT vs CRT 66% vs 83% (p = 0.05) AF vs AF 72% vs 77% (p = 0. 31) (Event: any failure) AF+C CF AF AF+C Year Probability P=0.01

Failure-Free Survival (3-year): 

Failure-Free Survival (3-year) CF This represents a significant interaction effect. What does it mean? Effects of CRT and AF are non-additive: Effect of CRT is enhanced in presence of AF and Effect of AF is enhanced in presence of CRT. Is it real? Is CRT useless without AF? Is AF useless without CRT?

Slide28: 

Hazard Ratio = .11 = 1/9  Nine-fold decrease in failure with AF + CRT together rather than separately. But the decrease is a product of three effects: RT + AF CRT + CF RT + CF ------------------------------------------------ CRT + AF CRT alone, HR  1.1 AF alone, HR  1.4 CRT + AF, HR  6 1.1  1.4  6 = 9, the total effect

Slide29: 

That is, the benefit due to CRT + AF being delivered in tandem vs. separately is due to: The apparent disadvantage of CRT + CF over RT + CF (1.1 increase); The apparent disadvantage of RT + AF over RT + CF (1.4 increase); and The advantage of CRT + AF over RT + CF (six-fold reduction): 1.1  1.4  6 = 9, the total interaction effect.

A clinically important question:: 

A clinically important question: Suppose we don’t believe the first two effects – we don’t believe that acceleration or adding chemotherapy can cause failures, and fix their hazard ratios at 1.0: What is the significance of the “restricted interaction effect”, the hazard ratio of 6? Answer: p = ????. Modest evidence that AF and CRT do seem to work better together.

Locoregional Failure-Free Rate (3-year): 

Locoregional Failure-Free Rate (3-year) RT vs CRT 81% vs 87% (p = 0.41) AF vs AF 83% vs 85% (p = 0. 63) Probability Year P=0.14 AF+C CF CF+C AF

Distant Failure-Free Rate (3-year): 

Distant Failure-Free Rate (3-year) RT vs CRT 79% vs 93% (p = 0.06) AF vs AF 85% vs 86% (p = 0. 55) Year Probability AF+C CF CF+C AF P=0.10

Disease Status at Last Assessment: 

CF+C 69 6 8 12 2 2 2 CF 71 5 7 14 - - 2 Disease Status at Last Assessment AF 67 6 4 19 - 2 2 AF+C 89 - - 5 - 7 -

Overall Survival : 

Overall Survival RT vs CRT 77% vs 87% (p = 0.41) AF vs AF 85% vs 79% (p = 0.73) Year Probability AF+C CF AF CF+C P=0.56

Progression-Free Survival : 

Progression-Free Survival (Event: failure or death) RT vs CRT 65% vs 80% (p = 0.19) AF vs AF 71% vs 74% (p = 0.34) Year Probability AF+C CF+C CF AF P=0.09

Late Toxicity (Grade >3): 

Late Toxicity (Grade >3) RT vs CRT 19% vs 33% (p = 0.08) AF vs AF 23% vs 28% (p = 0.32) Year Probability AF+C CF+C AF CF P=0.24

Late Toxicity (Grade >3): 

Late Toxicity (Grade >3)

Subset Analyses – Absolute Gain: 

Subset Analyses – Absolute Gain