Clinical study on guidance of the timing of adaptive replanning for cervical cancer based on ADC values combined with tumor volume regression

Technical Summary: Guidance of Adaptive Replanning Timing in Cervical Cancer Based on ADC Values and Tumor Volume Regression

Problem Statement
Concurrent chemoradiotherapy (CCRT) is the standard of care for locally advanced cervical cancer. While Intensity-Modulated Radiation Therapy (IMRT) is widely used, significant tumor regression often occurs during the treatment course, potentially rendering the initial treatment plan suboptimal. This anatomical change necessitates adaptive replanning to maintain target coverage and spare organs at risk (OARs). However, the optimal timing for initiating adaptive replanning remains uncertain. Indiscriminate early replanning increases workload and costs, while delayed replanning may compromise dose distribution. There is a clinical need for a practical imaging marker to identify which patients warrant early anatomical reassessment.

Methodology
The study was conducted in two distinct parts involving patients from the First Hospital of Qinhuangdao:

• Part I (Retrospective Cohort): Included 66 patients with locally advanced cervical cancer (FIGO IB3–IVA) treated between January 2022 and May 2024.

  • Imaging Protocol: Patients underwent MRI with Diffusion-Weighted Imaging (DWI) prior to treatment and at the 20th fraction of radiotherapy. Apparent Diffusion Coefficient (ADC) values were measured by a single radiologist.
  • Variables: The study analyzed the change in ADC ($\Delta$ADC), calculated as $(ADC_2 - ADC_1) / ADC_1 \times 100\%$. Tumor Volume Regression Rate (TVRR) was calculated to define "marked early regression" (median TVRR of the cohort was used as the cutoff: $\ge$ 85.56%).
  • Statistical Analysis: Logistic regression (univariable and multivariable) was used to assess the association between $\Delta$ADC and marked tumor volume regression. Receiver Operating Characteristic (ROC) analysis determined the discriminatory performance and optimal cutoff for $\Delta$ADC.

• Part II (Exploratory Dosimetric Simulation): Included 9 patients from a subsequent cohort (June–December 2024) who exhibited marked early imaging response ($\Delta$ADC $\ge$ 37.20%, based on Part I results).

  • Simulation Design: Three plans were generated for each patient:
    • Plan 1: Original treatment plan.
    • Plan 2: Replanning simulation at the 15th fraction.
    • Plan 3: Replanning simulation at the 20th fraction.
  • Comparison: Dosimetric parameters for the Planning Target Volume (PTV) and Organs at Risk (bladder and rectum), including $D_{2\%}$, $D_{mean}$, Conformity Index (CI), Homogeneity Index (HI), and volume metrics ($V_{40}$, $V_{50}$), were compared across the three time points using repeated-measures ANOVA or Friedman tests.

Key Results

• Tumor Response and ADC Changes: At the 20th fraction, median tumor volume decreased significantly from 27.54 cm³ to 4.33 cm³ ($P < 0.001$). Concurrently, ADC values increased significantly from $0.871 \pm 0.060$ to $1.191 \pm 0.083 \times 10^{-3}$ mm²/s ($P < 0.001$).
• Association with Regression:
  • Grouped Analysis: Patients with $\Delta$ADC $\ge$ 37.20% were significantly more likely to achieve marked tumor volume regression compared to those with lower $\Delta$ADC (29/33 vs. 4/33; $P < 0.001$).
  • Continuous Analysis: Each 5% increase in $\Delta$ADC was associated with a 6.09-fold higher odds of marked tumor volume regression (OR = 6.09, 95% CI 2.65–14.00, $P < 0.001$). This association remained significant after adjusting for baseline SCC-Ag and hemoglobin levels (OR = 6.12, $P < 0.001$).
  • Predictive Performance: ROC analysis yielded an Area Under the Curve (AUC) of 0.891, with an optimal cutoff of 37.20% (Sensitivity: 87.9%, Specificity: 87.9%).
• Dosimetric Simulation (Part II):
  • Anatomical Changes: Significant reductions were observed in Gross Tumor Volume (GTV), PTV, and bladder volume between Plan 1, Plan 2, and Plan 3.
  • Dosimetric Outcomes: Despite anatomical changes, there were no consistent dosimetric advantages for replanning at the 15th fraction (Plan 2) compared to the 20th fraction (Plan 3). PTV coverage (CI, HI, $D_{mean}$) and OAR sparing (bladder and rectal $V_{40}$, $V_{50}$) showed no statistically significant improvement with earlier replanning in this small cohort.

Key Contributions

1. Identification of a Predictive Marker: The study establishes that early changes in ADC ($\Delta$ADC) during CCRT are strongly associated with marked early tumor volume regression in cervical cancer.
2. Quantitative Threshold: It provides a preliminary cutoff value ($\Delta$ADC $\ge$ 37.20%) that may help identify patients with rapid tumor regression who could benefit from earlier anatomical reassessment.
3. Dosimetric Reality Check: Through exploratory simulation, the study demonstrates that while anatomical changes occur early, earlier replanning (at the 15th fraction) does not necessarily translate to superior dosimetric parameters compared to standard replanning (at the 20th fraction) in patients with marked response.

Significance and Claims
The paper concludes that $\Delta$ADC is a valuable imaging biomarker for identifying a subset of cervical cancer patients who warrant early anatomical reassessment during CCRT. However, the authors maintain a modest stance regarding the clinical implementation of early replanning. They explicitly state that while $\Delta$ADC can guide when to reassess anatomy, their dosimetric simulation did not prove that replanning at the 15th fraction offers a consistent dosimetric advantage over the 20th fraction. Therefore, the study suggests that $\Delta$ADC should be used to select patients for reassessment rather than as a mandate for universal early replanning. The authors emphasize that larger studies are required to confirm whether earlier replanning in these selected patients leads to tangible clinical benefits in toxicity or tumor control.