Purpose: In a prior study, the combination of (31)P magnetic resonance spectroscopy (MRS)-based intracellular pH (pHi) and T2 relaxation time was highly predictive of the pathologic complete response (pCR) rate in a small series of patients with soft tissue sarcomas (STSs) treated with thermoradiotherapy. Changes in the magnetic resonance metabolite ratios and pO(2) were related to the pCR rate. Hypoxia also correlated with a greater likelihood for the development of metastases. Because of the limited number of patients in the prior series, we initiated this study to determine whether the prior observations were repeatable and whether (31)P MRS lipid-related resonances were related to a propensity for metastasis.
Methods and materials: Patients with high-grade STSs were enrolled in an institutional review board-approved Phase II thermoradiotherapy trial. All tumors received daily external beam radiotherapy (1.8-2.0 Gy, five times weekly) to a total dose of 30-50 Gy. Hyperthermia followed radiotherapy by <1 h and was given two times weekly. Tumors were resected 4-6 weeks after radiotherapy completion. The MRS/MRI parameters included (31)P metabolite ratios, pHi, and T2 relaxation time. The median pO(2) and hypoxic fraction were determined using pO(2) histography. Comparisons between experimental endpoints and the pCR rate and metastasis-free and overall survival were made.
Results: Of 35 patients, 21 and 28 had reportable pretreatment MRS/MRI and pO(2) data, respectively. The cutpoints for a previously tested receiver operating curve for a pCR were T2 = 100 and pHi = 7.3. In the current series, few tumors fell below the cutpoints so validation was not possible. The phosphodiester (PDE)/inorganic phosphate (Pi) ratio and hypoxic fraction correlated inversely with the pCR rate in the current series (Spearman correlation coefficient -0.51, p = 0.017; odds ratio of percentage of necrosis > or =95% = 0.01 for a 1% increase in the hypoxic fraction; Wald p = 0.036). The pretreatment phosphomonoester (PME)/Pi ratio also correlated inversely with the pCR rate (odds ratio of percentage of necrosis > or =95% = 0.06 for pretreatment PME/Pi ratio >0.8 vs. < or =0.8, Wald p = 0.023). The pretreatment PME/PDE ratio correlated strongly with metastasis-free survival and overall survival (p = 0.012 and hazard ratio = 5.8, and p = 0.038 and hazard ratio = 6.75, respectively).
Conclusion: The dual parameter model containing pHi and T2 to predict the pCR in STSs treated with thermoradiotherapy was not verified. However, other parameters were statistically significant, including the PDE/Pi ratio and hypoxic fraction. These relationships may have interfered with our ability to obtain the pCR rate predicted by thermal doses achieved in these patients. The relationship between the PME/PDE ratio and metastasis-free and overall survival was provocative, but requires additional study to verify its predictive capability. Currently, 50% of all STS patients with high-grade tumors develop distant metastasis even when excellent local control is achieved. Parameters that could help select for patients who need adjuvant chemotherapy could have significant clinical benefit.