Purpose: The objective was to test whether tumor pH and 31P magnetic resonance spectroscopic end points were related to treatment outcome in pet canine patients with spontaneous soft tissue sarcomas treated with thermoradiotherapy.

Experimental Design: Forty-two dogs with evaluable 31P magnetic resonance spectroscopic end points and pH data were included in this study. Tumor variables (grade and volume), extracellular pH (pHe), T2 relaxation times, intracellular pH, and selected phosphometabolite ratios were examined for correlation with clinical outcome.

Results: From 39 dogs, pHe was a predictor of metastasis-free survival (MFS), with hazard ratio (HR, 0.29; P = 0.005) and overall survival (OS) with (HR, 0.36; P = 0.013). Tumor volume (>19 cm3) was related to MFS (HR, 2.14; P = 0.04), time to local failure (HR, 3.4; P = 0.025), and OS (HR, 2.27; P = 0.03). There was no association between T2 or intracellular pH and clinical outcome. Tumor grade (high versus low/intermediate) and phosphodiester/βATP ratio were identified as significant predictors for MFS, with (HR, 2.66; P = 0.009) and (HR, 0.75; P = 0.027), respectively, and as predictors of OS with (HR, 2.66; P = 0.009) and (HR, 0.76; P = 0.03), respectively. The phosphodiester/phosphocreatinine ratio predicted time to local failure (HR, 1.24; P = 0.017).

Conclusions: pHe was predictive of metastasis and OS in canine spontaneous sarcomas. To our knowledge, this is the first time that pHe has been shown to be predictive of clinical outcome. The results suggest that additional studies should be considered evaluating the prognostic significance of this variable. Phospholipid resonances, related to membrane metabolism, were related to clinical outcome, confirming recent results reported in human patients with soft tissue sarcomas treated with thermoradiotherapy.

Location, histologic grade, and tumor volume are three well-established predictors of clinical outcome for soft tissue sarcomas (STS) in humans and dogs (13). These factors, however, do not adequately identify patients who are at the highest risk for development of metastatic disease. Approximately half of the patients diagnosed with high-grade sarcomas will eventually develop metastases (4), but currently, there is no established test to distinguish which patients in this group are at highest risk. In addition, variation among pathologists leads to disagreement up to 25% of the time when staging sarcomas (4). For these reasons, imaging is being used more frequently to stage tumors in the clinic.

Magnetic resonance imaging (MRI) is used routinely to assess the characteristics and anatomic extent of tumors. Several groups have reported that contrast-enhanced MRI is of prognostic value for survival (5) and local changes in sarcomas in response to therapy (6). Additionally, T2 relaxation time and phosphorous magnetic resonance spectroscopy (P-31MRS) variables including intracellular pH (pHi) have been associated with the percentage of necrosis and duration of local tumor control following thermoradiotherapy in humans and dogs with STS (7, 8). P-31MRS metabolites reflect the energy status and rates of membrane turnover within the tumor. Previous reports have shown P-31MRS variables to be of prognostic importance in STS (711), bone sarcomas (12), head and neck tumors (13, 14), lymphomas (15, 16), and breast cancer (17).

Recently, we reported that phosphomonoester/phosphodiester (PDE), a ratio associated with lipid turnover, was predictive of metastasis in human patients with STS (9). Experimentally, metastasis has also been associated with the PDE/phosphomonoester ratio and extracellular pH (pHe) in a transgenic breast carcinoma model (18). These recent findings suggest phosphorous spectroscopy to be a valuable tool to obtain information related to malignant potential in STS, and more specifically, it indicates a relationship between metastasis and lipid biosynthetic profiles.

Although pHe can also be obtained with phosphorous spectroscopy under experimental conditions using contrast agents (i.e., 3-aminopropylphosphonate), this technique has not yet been introduced into the clinic (19). Needle probes can also be used to monitor pHe (20, 21). Although the probe technique is relatively inexpensive, its invasive nature has limited its introduction into the clinic. Microelectrode probe measurements have been shown to be prognostically important for tumor response and local control in prior thermoradiotherapy trials (2126), however, to date pHe has not been shown to be related to survival.

Indirect assessment of tumor acid/base status can be obtained by the measurement of lactate concentration. Lactate concentrations >8 μmol/g have been related to metastasis and worse overall survival (OS) in head and neck, colorectal, and cervical adenocarcinomas (2729). Lactate-lipid/creatinine ratios evaluated with proton (H1)-MRS were also related to survival in a population of 54 glioma patients (30).

In summary, prior studies have delineated the potential relationship between tumor acidity and/or membrane-related phospholipid metabolites and clinical prognosis. This report will focus on the evaluation of the importance of pHe and P-31MRS variables in predicting treatment outcome in pet dogs with spontaneous STS treated with thermoradiotherapy.

Patients and protocol. Forty-two dogs with spontaneous STS referred to the North Carolina State University Veterinary Medical Teaching Hospital were accrued onto a phase III clinical trial that included radiation therapy and hyperthermia from 1994 to 2001; the primary objective was to assess the effect of prospective application of controlled thermal dose on local tumor control following treatment. Treatment included 5 weeks of radiation therapy (2.25 Gy/d) using cobalt 60 photons and one hyperthermia treatment per week. Dogs were randomized into one of two thermal dose groups, in which there was a 10-fold difference in thermal dose between the two treatment arms. These animals were a subset of a larger phase III trial, the results of which has been reported elsewhere (31). It is important to note that for the overall trial, a significant difference in local tumor control was observed between the two treatment groups, with the high thermal dose group having a superior outcome. In the same trial, tumor volume and grade were shown to be prognostic for metastasis-free survival (MFS) and OS. This study was approved by the Institutional Animal Care and Use Committees from North Carolina State University and Duke University.

pHe. pHe was determined using interstitial needle electrodes (Microelectrode, Inc., Londonderry, NH; Agulian, Hamden, CT). The probes were precalibrated using standard buffer solutions and then were inserted aseptically into tumors through a small skin incision. pH measurements were obtained at 0.5 to 1 cm intervals along a track, at two to three different tumor locations (depending on the size of the tumor). Measurements were obtained before and after each MR procedure.

MRI/MRS. MRS and MRI studies were acquired before and 24 hours after the first hyperthermia treatment. Animals were anesthetized (isoflurane; Abbott Laboratories, North Chicago, IL) breathing 100% oxygen, and monitored closely during the MR procedure, which typically lasted >90 minutes. Images were acquired using a body coil, allowing the tumor to be in the center of the field of view with various CuSO4 standards for T2 calculations (40, 150, and 250 ms). Spectra were acquired with a surface dual frequency coil array and triphenyl phosphite standards solutions were placed next to the tumor for pulse calibration. Pulse sequence protocol and data analysis methods have been previously reported in detail (9).

Clinical assessment. All dogs were assessed after treatment for time to local failure, MFS, and OS as clinical end points. The time schedule for treatment and acquisition of data is graphically represented in Fig. 1. Follow-up was done at 3-month intervals for the first 2 years and every 6 months thereafter.

Fig. 1.

Graphical representation of treatment and physiologic variables obtained for this trial as a function of time.

Fig. 1.

Graphical representation of treatment and physiologic variables obtained for this trial as a function of time.

Close modal

Statistical analysis. Clinical outcome, survival variables (MFS, time to local failure, OS) and their relationship with MRI/MRS variables and pH, were evaluated using Kaplan-Meier graphs, log-rank tests, and Cox proportional hazards analysis. Cut points were determined by using medians (or 0 for a pre-versus post-change variable) of respective continuous variables or by using values from previously reported studies. As a secondary analysis, a multivariable regression analysis was done in order to discern possible interactions among significant variables. To calculate HR for a continuous predictor in a Cox proportional hazards model setting, half of the interquartile range (IQR/2) of the corresponding predictor was used as the increment of the predictor for HR.

Patients. A total of 42 dogs had concurrent pHe and MRS measurements. Of these, three were not included in the analysis because of poor signal to noise from the MRS data, or failure to converge during the fitting algorithm. This left 39 dogs that had data of sufficient quality to be evaluable (93%). From 21 male dogs, 17 were castrated, and from 18 female dogs, 17 were spayed. Most of the dogs were elderly with a median age of 10 years (Table 1).

Table 1.

Patient, endpoints, and treatment characteristics

SexTumor histologyTumor gradeAge (y)Tumor volume (cc)CEM43°C T90 (HT1)T90 (HT1)T2 (ms)pHi (pre)pHi (post)PDE/ PCrPDE/βATPpHe (pre)pHe (post)
Low thermal dose               
    1 SF HPC Low 14 210.6 1.52 41.2 113.71 7.18 — 2.11 1.19 7.06 — 
    2 SF HPC Low 15 13.4 1.01 39.2 100.17 7.23 7.17 0.89 2.68 6.61 6.68 
    3 SF MYX Low 10 11 1.79 40.35 92.98 — — — — 6.47 7.00 
    4 CM FSA Low 1.3 0.30 39.9 71.84 — — — — — — 
    5 SF HPC Low 10 2.5 0.58 40.2 59.33 — — — — — — 
    6 CM MYX Low 84.6 2.56 41.8 151.13 6.77  5.20 3.38 6.85 7.00 
    7 SF SA Low 3.4 0.62 40.3 3,617.42 — — — — 7.00  
    8 SF NFS Low 10 15.4 0.75 40.08 1,030.58 7.30 6.44 6.12 9.95 7.01 7.01 
    9 CM HPC Low 64.25 0.63 40.1 2,552.45 7.37 7.53 0.57 5.22 7.09 7.07 
    10 HPC Low 4.38 0.64 37.74 — 7.62 7.41 1.02 0.94 7.23 7.26 
    11 SF HPC Low 10 33 0.83 39.6 71.20 7.27 7.77 2.19 1.42 7.28 6.85 
    12 SF HPC Low 13 5.1 1.26 40.3 — 7.07 6.93 0.18 3.92 7.32 7.28 
    13 CM HPC Low 257.3 1.12 40.6 — 7.09 — 1.23 2.55 6.95 — 
    14 SA Low 13 18.8 2.16 39.8 55.21 7.26 — 0.50 2.14 — — 
    15 SF SA Low 38.9 0.67 39.75 80.81 7.18 — 0.25 0.89 7.11 7.16 
    16 CM HPC High 125.4 0.96 39.82 92.76 7.19 7.36 0.17 1.03 7.00 6.94 
    17 CM FSA High 10 40.4 0.91 37.9 77.98 — — — — 6.88 6.91 
    18 CM FSA High 11 28.8 0.22 38.82 72.82 7.33 7.42 0.47 1.36 7.12 7.01 
    19 SF HPC High 15.7 0.47 39.92 73.53 — — — — — — 
    20 CM HPC High 15 13.2 0.64 40 — 7.37 — 0.20 0.64 7.06 7.03 
    21 CM SA High 10 109.03 1.80 40.2 68.71 7.02 7.41 0.53 0.77 — 6.98 
High thermal dose               
    22 CM HPC Low 7.3 12.62 40.7 7,360.84 7.04 — — 4.65 7.05 7.11 
    23 SF HPC Low 14 81.9 12.46 40.9 147.98 7.21 — 15.08 2.19 7.09 — 
    24 CM MYX Low 10 2.7 10.09 40.8 63.83 — — — — — 7.27 
    25 CM FSA Low 11 12.7 6.81 40.7 108.75 — — — — 6.72 — 
    26 FSA Low 120 2.87 40.53 128.34 7.25 7.09 1.59 2.98 6.89 6.61 
    27 CM FSA Low 2.3 8.44 41.66 5.53 7.39  0.20 0.78 — — 
    28 SF HPC Low 12 122.5 1.55 40.17 61.04 7.58 7.13 5.39 1.22 6.93 6.59 
    29 CM MYX Low 15.7 8.04 40.9 4,952.28 7.28 7.20 0.23 1.74 7.32 6.97 
    30 CM HPC Low 11 66.69 12.07 50 79.40 7.28 7.51 3.86 4.11 6.65 7.14 
    31 CM NFS Low 11 13.9 10.97 39.1 — 7.24 — 0.86 3.22 7.25 7.12 
    32 SF FSA Low 33.2 0.49 38.29 — 7.69 — 3.19 7.03 7.24 7.06 
    33 SF HPC Intermediate 12 19.1 14.45 40.7 107.45 7.27  0.11 0.27 6.48 6.65 
    34 SF FSA High 14 80.8 8.51 40.3 84.96 7.26 — 2.04 — — — 
    35 CM HPC High 11 7.2 7.15 41 53.78 7.31  1.04 3.31 7.21 — 
    36 FSA High 45.3 1.77 40.02 215.48 7.19 7.17 0.21 1.04 6.85 7.08 
    37 HPC High 11 3.7 5.38 40.9 28.44 7.34  0.28 1.32 — — 
    38 SF HPC High 13 64.5 9.94 44.12 68.34 — — — — 7.04 6.78 
    39 SF HPC High 13 66.4 7.53 41.7 130.37 7.58 7.45 1.22 3.74 7.23 6.91 
SexTumor histologyTumor gradeAge (y)Tumor volume (cc)CEM43°C T90 (HT1)T90 (HT1)T2 (ms)pHi (pre)pHi (post)PDE/ PCrPDE/βATPpHe (pre)pHe (post)
Low thermal dose               
    1 SF HPC Low 14 210.6 1.52 41.2 113.71 7.18 — 2.11 1.19 7.06 — 
    2 SF HPC Low 15 13.4 1.01 39.2 100.17 7.23 7.17 0.89 2.68 6.61 6.68 
    3 SF MYX Low 10 11 1.79 40.35 92.98 — — — — 6.47 7.00 
    4 CM FSA Low 1.3 0.30 39.9 71.84 — — — — — — 
    5 SF HPC Low 10 2.5 0.58 40.2 59.33 — — — — — — 
    6 CM MYX Low 84.6 2.56 41.8 151.13 6.77  5.20 3.38 6.85 7.00 
    7 SF SA Low 3.4 0.62 40.3 3,617.42 — — — — 7.00  
    8 SF NFS Low 10 15.4 0.75 40.08 1,030.58 7.30 6.44 6.12 9.95 7.01 7.01 
    9 CM HPC Low 64.25 0.63 40.1 2,552.45 7.37 7.53 0.57 5.22 7.09 7.07 
    10 HPC Low 4.38 0.64 37.74 — 7.62 7.41 1.02 0.94 7.23 7.26 
    11 SF HPC Low 10 33 0.83 39.6 71.20 7.27 7.77 2.19 1.42 7.28 6.85 
    12 SF HPC Low 13 5.1 1.26 40.3 — 7.07 6.93 0.18 3.92 7.32 7.28 
    13 CM HPC Low 257.3 1.12 40.6 — 7.09 — 1.23 2.55 6.95 — 
    14 SA Low 13 18.8 2.16 39.8 55.21 7.26 — 0.50 2.14 — — 
    15 SF SA Low 38.9 0.67 39.75 80.81 7.18 — 0.25 0.89 7.11 7.16 
    16 CM HPC High 125.4 0.96 39.82 92.76 7.19 7.36 0.17 1.03 7.00 6.94 
    17 CM FSA High 10 40.4 0.91 37.9 77.98 — — — — 6.88 6.91 
    18 CM FSA High 11 28.8 0.22 38.82 72.82 7.33 7.42 0.47 1.36 7.12 7.01 
    19 SF HPC High 15.7 0.47 39.92 73.53 — — — — — — 
    20 CM HPC High 15 13.2 0.64 40 — 7.37 — 0.20 0.64 7.06 7.03 
    21 CM SA High 10 109.03 1.80 40.2 68.71 7.02 7.41 0.53 0.77 — 6.98 
High thermal dose               
    22 CM HPC Low 7.3 12.62 40.7 7,360.84 7.04 — — 4.65 7.05 7.11 
    23 SF HPC Low 14 81.9 12.46 40.9 147.98 7.21 — 15.08 2.19 7.09 — 
    24 CM MYX Low 10 2.7 10.09 40.8 63.83 — — — — — 7.27 
    25 CM FSA Low 11 12.7 6.81 40.7 108.75 — — — — 6.72 — 
    26 FSA Low 120 2.87 40.53 128.34 7.25 7.09 1.59 2.98 6.89 6.61 
    27 CM FSA Low 2.3 8.44 41.66 5.53 7.39  0.20 0.78 — — 
    28 SF HPC Low 12 122.5 1.55 40.17 61.04 7.58 7.13 5.39 1.22 6.93 6.59 
    29 CM MYX Low 15.7 8.04 40.9 4,952.28 7.28 7.20 0.23 1.74 7.32 6.97 
    30 CM HPC Low 11 66.69 12.07 50 79.40 7.28 7.51 3.86 4.11 6.65 7.14 
    31 CM NFS Low 11 13.9 10.97 39.1 — 7.24 — 0.86 3.22 7.25 7.12 
    32 SF FSA Low 33.2 0.49 38.29 — 7.69 — 3.19 7.03 7.24 7.06 
    33 SF HPC Intermediate 12 19.1 14.45 40.7 107.45 7.27  0.11 0.27 6.48 6.65 
    34 SF FSA High 14 80.8 8.51 40.3 84.96 7.26 — 2.04 — — — 
    35 CM HPC High 11 7.2 7.15 41 53.78 7.31  1.04 3.31 7.21 — 
    36 FSA High 45.3 1.77 40.02 215.48 7.19 7.17 0.21 1.04 6.85 7.08 
    37 HPC High 11 3.7 5.38 40.9 28.44 7.34  0.28 1.32 — — 
    38 SF HPC High 13 64.5 9.94 44.12 68.34 — — — — 7.04 6.78 
    39 SF HPC High 13 66.4 7.53 41.7 130.37 7.58 7.45 1.22 3.74 7.23 6.91 

Abbreviations: T90, temperature at 10th percentile of the distribution; CEM43°C T90, cumulative equivalent minutes at 43°C for T90; HT1, first heat; T2, T2 relaxativity time; M, male; MC, male castrated; F, female; SF, spayed female; HPC, hemangiopericytoma; FSA, fibrosarcoma; MYX, myxoma; SA, undifferenciated sarcoma; NFS, neurofribrosarcoma.

Tumors. Hemangiopericytoma was the most commonly observed tumor (n = 20) followed by fibrosarcoma (n = 9), myxoma (n = 4), undifferentiated sarcoma (n = 4), and neurofibrosarcoma (n = 2). Most of the tumors were low grade (n = 26). The remaining tumors were intermediate (n = 1) or high grade (n = 12; Table 1). Tumor grade (high versus low/intermediate) was related to MFS and OS with the same hazard ratio (HR) for both end points (HR, 2.66; P = 0.009). Tumor volume (median, 19.1; IQR, 59.39 cm3) was related to MFS (Fig. 2; HR, 2.14; P = 0.04), OS (HR, 2.27; P = 0.03), and time to local failure (HR, 3.4; P = 0.025; Tables 2 and 3).

Fig. 2.

Kaplan-Meier plot representing MFS for canine STS with tumor volume above and below the median cut point.

Fig. 2.

Kaplan-Meier plot representing MFS for canine STS with tumor volume above and below the median cut point.

Close modal
Table 2.

Cox proportional hazards ratio for physiological parameters on clinical outcome variables

Type of variable or cut pointnMFS
Time to local failure
OS
HR (95% CI)Wald P valueHR (95% CI)Wald P valueHR (95% CI)Wald P value
Tumor volume (cc) >19 versus ≤19 39 2.14 (1.03-4.42) 0.041 3.4 (1.17-9.9) 0.025 2.27 (1.08-4.8) 0.031 
Tumor grade High versus low/intermediate 39 2.66 (1.27-5.54) 0.009 2.3 (0.87-6.13) 0.095 2.66 (1.28-5.56) 0.009 
T2 (ms) >100 versus ≤100 33 1.10 (0.51-2.36) 0.812 0.80 (0.28-2.26) 0.675 1.08 (0.50-2.30) 0.850 
PDE/PCr Continuous 29 0.95 (0.81-1.10) 0.478 1.24 (1.04-1.48) 0.017 0.95 (0.82-1.10) 0.505 
PDE/βATP Continuous 29 0.75 (0.59-0.97) 0.027 0.87 (0.65-1.17) 0.347 0.76 (0.59-0.97) 0.030 
pHi >7.26 versus ≤7.26 30 0.54 (0.24-1.25) 0.150 0.49 (0.16-1.46) 0.199 0.58 (0.25-1.33) 0.199 
pHi (post-pre) >−0.06 versus ≤−0.06 15 1.44 (0.45-4.60) 0.538 7.70 (0.89-66.2) 0.063 1.41 (0.44-4.50) 0.565 
pHe >7 versus ≤7 30 0.29 (0.12-0.69) 0.005 0.84 (0.27-2.60) 0.767 0.36 (0.16-0.81) 0.013 
pHe (post-pre) >0 versus ≤0 24 2.87 (1.09-7.58) 0.034 1.59 (0.46-5.46) 0.465 2.54 (0.96-6.74) 0.061 
phi/pHe (gradient) >0.21 versus ≤0.21 25 1.15 (0.47-2.80) 0.757 0.77 (0.23-2.65) 0.681 1.19 (0.49-2.90) 0.697 
Type of variable or cut pointnMFS
Time to local failure
OS
HR (95% CI)Wald P valueHR (95% CI)Wald P valueHR (95% CI)Wald P value
Tumor volume (cc) >19 versus ≤19 39 2.14 (1.03-4.42) 0.041 3.4 (1.17-9.9) 0.025 2.27 (1.08-4.8) 0.031 
Tumor grade High versus low/intermediate 39 2.66 (1.27-5.54) 0.009 2.3 (0.87-6.13) 0.095 2.66 (1.28-5.56) 0.009 
T2 (ms) >100 versus ≤100 33 1.10 (0.51-2.36) 0.812 0.80 (0.28-2.26) 0.675 1.08 (0.50-2.30) 0.850 
PDE/PCr Continuous 29 0.95 (0.81-1.10) 0.478 1.24 (1.04-1.48) 0.017 0.95 (0.82-1.10) 0.505 
PDE/βATP Continuous 29 0.75 (0.59-0.97) 0.027 0.87 (0.65-1.17) 0.347 0.76 (0.59-0.97) 0.030 
pHi >7.26 versus ≤7.26 30 0.54 (0.24-1.25) 0.150 0.49 (0.16-1.46) 0.199 0.58 (0.25-1.33) 0.199 
pHi (post-pre) >−0.06 versus ≤−0.06 15 1.44 (0.45-4.60) 0.538 7.70 (0.89-66.2) 0.063 1.41 (0.44-4.50) 0.565 
pHe >7 versus ≤7 30 0.29 (0.12-0.69) 0.005 0.84 (0.27-2.60) 0.767 0.36 (0.16-0.81) 0.013 
pHe (post-pre) >0 versus ≤0 24 2.87 (1.09-7.58) 0.034 1.59 (0.46-5.46) 0.465 2.54 (0.96-6.74) 0.061 
phi/pHe (gradient) >0.21 versus ≤0.21 25 1.15 (0.47-2.80) 0.757 0.77 (0.23-2.65) 0.681 1.19 (0.49-2.90) 0.697 

NOTE: T2, T2 relaxation time.

Table 3.

Kaplan-Meier and log-rank test results on the significant variables found on this study

VariablesCut pointMFS
Time to local failure
OS
Median95% CI (lower)95% CI (upper)Log rankMedian95% CI (lower)95% CI (upper)Log rankMedian95% CI (lower)95% CI (upper)Log rank
pHe ≤7 1.3 0.31 2.21 0.0029 — 0.26 — 0.7681 1.3 0.54 2.21 0.0102 
 >7 3.29 1.31 4.48  2.06 1.02 —  3.29 1.33 4.48  
Change in pHE ≤0 2.15 1.3 4.48 0.0267 1.62 0.72 — 0.4611 2.15 1.3 4.48 0.0525 
 >0 0.63 0.25 2.3  — 0.21 —  1.22 0.54 2.3  
PDE/βATP ≤2.14 1.17 0.37 1.31 0.1028 0.59 0.37 — 0.0797 1.25 0.49 1.7 0.1203 
 >2.14 2.93 1.83 4.15  — 1.47 —  2.93 1.83 4.15  
Tumor volume ≤19 3.39 1.83 4.48 0.0367 — 1.62 — 0.0179 3.75 1.83 4.62 0.0269 
 >19 1.31 0.49 2.66  0.72 0.37 —  1.32 0.57 2.66  
Tumor grade Low 2.93 1.76 3.84 0.0069 4.08 1.47 — 0.0866 2.93 1.76 4.15 0.0068 
 Intermediate/high 1.22 0.37 1.33  0.52 0.35 —  1.22 0.54 1.33  
VariablesCut pointMFS
Time to local failure
OS
Median95% CI (lower)95% CI (upper)Log rankMedian95% CI (lower)95% CI (upper)Log rankMedian95% CI (lower)95% CI (upper)Log rank
pHe ≤7 1.3 0.31 2.21 0.0029 — 0.26 — 0.7681 1.3 0.54 2.21 0.0102 
 >7 3.29 1.31 4.48  2.06 1.02 —  3.29 1.33 4.48  
Change in pHE ≤0 2.15 1.3 4.48 0.0267 1.62 0.72 — 0.4611 2.15 1.3 4.48 0.0525 
 >0 0.63 0.25 2.3  — 0.21 —  1.22 0.54 2.3  
PDE/βATP ≤2.14 1.17 0.37 1.31 0.1028 0.59 0.37 — 0.0797 1.25 0.49 1.7 0.1203 
 >2.14 2.93 1.83 4.15  — 1.47 —  2.93 1.83 4.15  
Tumor volume ≤19 3.39 1.83 4.48 0.0367 — 1.62 — 0.0179 3.75 1.83 4.62 0.0269 
 >19 1.31 0.49 2.66  0.72 0.37 —  1.32 0.57 2.66  
Tumor grade Low 2.93 1.76 3.84 0.0069 4.08 1.47 — 0.0866 2.93 1.76 4.15 0.0068 
 Intermediate/high 1.22 0.37 1.33  0.52 0.35 —  1.22 0.54 1.33  

Thermal dose. Thermal dose was measured with the CEM43°C T90 (cumulative equivalent minutes at 43°C with temperatures above the 10th percentile of the distribution of temperatures), which is a well-accepted way of measuring heat delivery (32). The low thermal dose group (median, 2.87 CEM43°C T90; IQR, 1.28) was significantly lower than in the high thermal dose group (median, 47.62 CEM43°C T90; IQR, 14.7) P < 0.0001 by Wilcoxon-rank sum test. There was a trend toward longer duration of local control in the high thermal dose group, similar to that found to be significant in the whole trial (31).

pHe. The median pHe was 7.0, and this value was used as a cut point for statistical analysis. There was no difference in pHe between the low and high thermal dose groups. pHe was a highly significant predictor of MFS (HR, 0.29; P = 0.005) and OS (HR, 0.36; P = 0.013; Table 2; Fig. 3A). The prognosis was worse overall for tumors with relatively acidic pHe. The change in pHe after the first hyperthermia treatment was also significantly related to MFS (HR, 2.87; P = 0.034). Tumors that acidified (posttreatment) had a better prognosis than tumors that became more alkaline (Fig. 3B). A possible relationship between pHe and tumor volume that was discarded based on multivariable regression analysis; however, was inconclusive when relating pHe to tumor grade (Fig. 4A).

Fig. 3.

Kaplan-Meier plot representing MFS for pHe (A) and relative change in tumor pHe 24 hours post-hyperthermia treatment (B) above and below the median and 0, respectively.

Fig. 3.

Kaplan-Meier plot representing MFS for pHe (A) and relative change in tumor pHe 24 hours post-hyperthermia treatment (B) above and below the median and 0, respectively.

Close modal
Fig. 4.

Relationship between tumor volume (A) and relative change in tumor pHe 24 hours posttreatment (B) as a function of baseline pHe (pre).

Fig. 4.

Relationship between tumor volume (A) and relative change in tumor pHe 24 hours posttreatment (B) as a function of baseline pHe (pre).

Close modal

MRI/MRS. T2 measurements of 33 dogs (median, 85 ms; IQR, 61.7 ms) obtained prior to treatment were not related to pHi or pHe, and were not predictive of clinical outcome (P > 0.05; data not shown). The remaining six animals were unevaluable for T2 because of errors in placement of the T2 standards such that they were not visible in tumor-containing slices.

Figure 5 depicts an example of a P-31 MR spectrum obtained from a fibrosarcoma. Several P-31 MRS variables were examined for their relationship with outcome. Of these, the PDE/βATP ratio was significantly associated with MFS (HR, 0.75; P = 0.027) and OS (HR, 0.76; P = 0.03). When dichotomized (>2.14 versus <2.14), the PDE/βATP ratio was not significant, but an important median survival difference of 20 months between both arms was found. The PDE/phosphocreatinine (PCr) ratio was also moderately predictive of tumor local control probability (HR, 1.24; P = 0.017; Table 2; Fig. 6).

Fig. 5.

P-31MRS spectrum from a hemangiopericytoma located above the right elbow. Three-dimensional chemical shift imaging voxels depicted, along with an average spectrum from two tumor-containing voxels.

Fig. 5.

P-31MRS spectrum from a hemangiopericytoma located above the right elbow. Three-dimensional chemical shift imaging voxels depicted, along with an average spectrum from two tumor-containing voxels.

Close modal
Fig. 6.

Kaplan-Meier plot representing MFS for PDE/βATP for canine STS above and below the median cut point.

Fig. 6.

Kaplan-Meier plot representing MFS for PDE/βATP for canine STS above and below the median cut point.

Close modal

Pretreatment intracellular pH (median, 7.26; IQR, 0.15), its difference posttreatment (post-pre), and the intracellular/extracellular gradient were not related to clinical outcome.

Although this trial confirms the importance of tumor volume and tumor grade as factors influencing survival in dogs with STS following thermoradiotherapy, the association of pHe effects and P-31 metabolites with metastasis and survival is provocative. Here, we can see that tumors with acidic extracellular pretreatment pH, based on needle electrode measurements, tend to metastasize more than three times faster than nonacidic tumors. Also, tumors that acidify 24 hours post-first heat, as well as tumors with high PDE/βATP ratios, are less likely to exhibit metastasis. These clinical results show the hypothesized relationship between acidity, lipid metabolism, and tumor metastasis (18).

Relationship between pHe and treatment outcome. Our study shows for the first time the influence of pHe in predicting MFS and OS. STS with pHe <7.0, had a greater probability to metastasize, compared with tumors with pHe >7.0. pHe has been evaluated for its relationship with the likelihood of achieving complete tumor response in two clinical trials involving the use of thermoradiotherapy for recurrent or metastatic superficial tumors in prior studies (22, 23). In a report of 26 patients with heterogeneous histologic types enrolled in a phase II trial, a low average pHe (pHe = 6.88) was observed in patients who achieved complete response, whereas a higher pHe average (pHe = 7.24) was observed in patients who achieved either partial or no response (22). In contrast, the opposite result was observed in 49 patients with predominantly chest wall recurrences of breast cancer. Patients who achieved a complete response had higher average pH (pH 7.36) compared with those who achieved either no or partial response (pH 7.16; ref. 23). Possible explanations for the differences between these trials may be methodologic or may be related to the particular histologic types being evaluated. The measurement of pH by needle electrodes is not technically difficult, but it is subject to error. It is important to wait until the electrode reaches steady state after being inserted into tissue, and calibrations prior to use are highly desirable to verify that the probes are functioning properly. Nevertheless, the microelectrode is a well-accepted instrument to determine pHe in vivo.

There is a rationale for hypothesizing that pHe might be involved in tumor progression. Tumor cells exposed to low pHe (6.8) activate proteases (cathepsin B and metalloproteinases) that are involved in extracellular matrix degradation (33). Melanoma cells have been reported to increase the secretion of type IV collagenase at low pHe (34, 35). Mathematical models have predicted that acidification may be a mechanism for tumor cell invasion (36) by activating cysteine proteinases that are important during metastasis (37).

Provocative preclinical data have been published relating to the effect of pHe on metastatic behavior. Bhujwalla et al. examined the differences in extracellular and intracellular pH between two transgenic tumor lines that differed in expression of nm-23H (nonmetastatic protein 23H; ref. 18). Using P-31MRS, they measured both pH values, where pHe was measured using a contrast agent. pHe for highly metastatic tumors was 6.8 compared with 7.13 (average) for wild-type controls. This is indirect evidence for a connection between tumor acidosis and metastatic behavior. Low pH is a characteristic of hypoxic tumors (38, 39), which are radioresistant and typically have poor prognosis (40).

The relationship between tumor metastasis and the increase in pHe, seen 24 hours post-hyperthermia is interesting, but difficult to explain. It is known that high temperatures can acidify the microenvironment up to 24 hours after heating (41). However, we did not observe a significant relationship between the thermal dose achieved and the change in pH (data not shown). In addition, we could see that tumors that were acidic at the beginning became alkaline post-therapy and vice versa (Fig. 4B). Thus, the change in pHe and pretreatment pHe values are probably related and is telling us similar information.

Correlative analysis was done to investigate whether pHe could be an independent predictor of metastasis or OS. We found no relation between pHe (or its gradient) and tumor volume. In previous preclinical rodent studies, tumor pH values decrease whereas tumor volume (or mass) increases (38, 42, 43). For example, large ulcerating mammary carcinomas implanted in rodents are more acidic (0.6-0.8 units) than smaller ones (38) and a decrease of 0.1 on average pHe levels is seen in tumors growing from 1 to 2.5 g and from 4 to 6 g (42). This phenomena is very common in experimental models with fast-growing tumors, but does not seem to be an important relationship in this series of spontaneous canine sarcomas.

Due to the low number of high-grade tumors, the relationship between tumor grade and pHe in the multivariable regression analysis was inconclusive. It is possible that there may be such an association because more aggressive tumors tend to have higher growth rates and may have an increased glycolytic metabolism. Lactate concentration in tumors has been reported to be prognostically important in several small clinical series including head and neck, colorectal, and cervix cancer (2729). In head and neck cancer, high lactate levels were associated with an increased likelihood for metastases and reduced survival, but not local tumor control (29). None of the series reported thus far are sufficiently large to perform full multivariate analysis. It will be important to determine whether pH or lactate levels are independently predictive of outcome, compared with other known prognostic factors.

Relationship between pHi and treatment outcome. As previously reported (44), we found that spontaneous canine tumors tend to be alkaline intracellularly (median, 7.26), whereas being acidic extracellularly (median, 7.0). Contrary to pHe, we found that pHi was not significantly related to clinical outcome. In a prior report, we showed that high pHi values obtained before treatment were predictive of pathologic complete response rate in human STS, particularly when this variable was combined with T2 relaxation times (8). In a recent report of a follow-up trial, neither of these variables were predictive of pathologic complete response (9). There have been significant technologic improvements in P-31MRS between these two reports, with the transition from one-dimensional to three-dimensional chemical shift imaging. The practice in the current study of eliminating voxels with significant normal tissue contamination may be particularly important here.

P-31 metabolites and treatment outcome. The PDE/βATP ratio was significant as a continuous variable in predicting MFS and OS. When dichotomized at its median of ratio of 2.14, there is a 21-month difference in median times to metastasis and survival. This canine population was elderly (10 years), and that the average life span for a dog is around 12 years. In this population, therefore, a minimum of 20-month difference in median times to metastasis and survival represents an approximately 20% improvement in life span, which is substantial. If this translates to equivalent differences in survival in humans, it may be a very important prognostic factor for sarcomas. We previously reported changes in the inverse of this ratio to be correlated with pathologic complete response and duration of local control in human and canine patients with STS, respectively (45).

The PDE/PCr ratio was the only ratio that predicted local failure in our study and was also previously reported (in conjunction with the phosphomonoester/PDE ratio) to be related to MFS and OS in 35 human patients with STS treated with thermoradiotherapy (9). An increase in PDEs seems to be the main reason for an increase in PDE/PCr (as well as PDE/βATP) because PCr levels are always low in tumor tissues. Elevated values of PDEs are associated with an increase in phospholipid catabolism, elevated cell membrane breakdown, and necrosis in tumors (46, 47). In a prior report, glycerophosphocoline and glycerophosphoethanolamine (PDE) levels were found to be elevated in a low metastatic cell strain suggesting phospholipid signaling mechanisms (18). Although our results are consistent with this prior study, further work will need to be done to clarify the effects of PDE metabolism on tumor progression and invasion.

In summary, we describe the predictive properties of pHe in STS treated with thermoradiotherapy. Higher pretreatment pHe and tumors that did not become alkaline after first hyperthermia treatment are characterized by longer MFS and OS times. Although this concept has been shown in experimental transplantable tumor models, to our knowledge, this is the first time pHe has been identified as a significant prognostic factor in spontaneous tumors. It is still necessary to confirm these findings in a larger population and to evaluate the relationship with tumor grade. The development of noninvasive methods, especially in MRI and spectroscopy, to noninvasively determine pHe in tumors would be extremely helpful in this regard (19, 4851). The increase in PDE ratios, seen in PDE/βATP and PDE/PCr association with clinical outcome may be explained by elevations in lipid metabolism rates.

Grant support: NIH grant PO1CA42745 from the Department of Health and Human Services.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Forrest LJ, Chun R, Adams WM, Cooley AJ, Vail DM. Postoperative radiotherapy for canine soft tissue sarcoma.
J Vet Intern Med
2000
;
14
:
578
–82.
2
Kuntz CA, Dernell WS, Powers BE, Devitt C, Straw RC, Withrow SJ. Prognostic factors for surgical treatment of soft-tissue sarcomas in dogs: 75 cases (1986–1996).
J Am Vet Med Assoc
1997
;
211
:
1147
–51.
3
Zagars GK, Ballo MT, Pisters PW, et al. Prognostic factors for patients with localized soft-tissue sarcoma treated with conservation surgery and radiation therapy: an analysis of 1225 patients.
Cancer
2003
;
97
:
2530
–43.
4
Pisters PW, O'Sullivan B. Soft tissue sarcomas. In: Kufe D, Pollock R, Weichselbaum R, editors. Cancer medicine. 6th ed. Hamilton: BC Decker; 2003. p. 2049–76.
5
Miller SL, Hoffer FA, Reddick WE, et al. Tumor volume or dynamic contrast-enhanced MRI for prediction of clinical outcome of Ewing sarcoma family of tumors.
Pediatr Radiol
2001
;
31
:
518
–23.
6
Preda A, Wielopolski PA, Ten Hagen TL, et al. Dynamic contrast-enhanced MRI using macromolecular contrast media for monitoring the response to isolated limb perfusion in experimental soft-tissue sarcomas.
MAGMA
2004
;
17
:
296
–302.
7
Dewhirst MW, Sostman HD, Leopold KA, et al. Soft-tissue sarcomas: MR imaging and MR spectroscopy for prognosis and therapy monitoring. Work in progress.
Radiology
1990
;
174
:
847
–53.
8
Sostman HD, Prescott DM, Dewhirst MW, et al. MR imaging and spectroscopy for prognostic evaluation in soft-tissue sarcomas.
Radiology
1994
;
190
:
269
–75.
9
Dewhirst MW, Poulson JM, Yu D, et al. Relation between pO(2), (31)P magnetic resonance spectroscopy parameters and treatment outcome in patients with high-grade soft tissue sarcomas treated with thermoradiotherapy.
Int J Radiat Oncol Biol Phys
2005
;
61
:
480
–91.
10
Koutcher JA, Ballon D, Graham M, et al. 31P NMR spectra of extremity sarcomas: diversity of metabolic profiles and changes in response to chemotherapy.
Magn Reson Med
1990
;
16
:
19
–34.
11
Sijens PE, Eggermont AM, van Dijk PV, Oudkerk M. 31P magnetic resonance spectroscopy as predictor of clinical response in human extremity sarcomas treated by single dose TNF-α + melphalan isolated limb perfusion.
NMR Biomed
1995
;
8
:
215
–24.
12
Zakian KL, Shukla-Dave A, Meyers P, et al. Identification of prognostic markers in bone sarcomas using proton-decoupled phosphorus magnetic resonance spectroscopy.
Cancer Res
2003
;
63
:
9042
–7.
13
Maldonado X, Alonso J, Giralt J, et al. 31Phosphorus magnetic resonance spectroscopy in the assessment of head and neck tumors.
Int J Radiat Oncol Biol Phys
1998
;
40
:
309
–12.
14
Shukla-Dave A, Poptani H, Loevner LA, et al. Prediction of treatment response of head and neck cancers with P-31 MR spectroscopy from pretreatment relative phosphomonoester levels.
Acad Radiol
2002
;
9
:
688
–94.
15
Smith SR, Martin PA, Davies JM, Edwards RH, Stevens AN. The assessment of treatment response in non-Hodgkin's lymphoma by image guided 31P magnetic resonance spectroscopy.
Br J Cancer
1990
;
61
:
485
–90.
16
Arias-Mendoza F, Smith MR, Brown TR. Predicting treatment response in non-Hodgkin's lymphoma from the pretreatment tumor content of phosphoethanolamine plus phosphocholine.
Acad Radiol
2004
;
11
:
368
–76.
17
Twelves CJ, Porter DA, Lowry M, et al. Phosphorus-31 metabolism of post-menopausal breast cancer studied in vivo by magnetic resonance spectroscopy.
Br J Cancer
1994
;
69
:
1151
–6.
18
Bhujwalla ZM, Aboagye EO, Gillies RJ, Chacko VP, Mendola CE, Backer JM. Nm23-transfected MDA-MB-435 human breast carcinoma cells form tumors with altered phospholipid metabolism and pH: a 31P nuclear magnetic resonance study in vivo and in vitro.
Magn Reson Med
1999
;
41
:
897
–903.
19
Gillies RJ, Liu Z, Bhujwalla Z. 31P-MRS measurements of extracellular pH of tumors using 3-aminopropylphosphonate.
Am J Physiol
1994
;
267
:
C195
–203.
20
Evelhoch JL. pH and therapy of human cancers [discussion 4, 152–3].
Novartis Found Symp
2001
;
240
:
68
–80.
21
Wike-Hooley JL, Haveman J, Reinhold HS. The relevance of tumour pH to the treatment of malignant disease.
Radiother Oncol
1984
;
2
:
343
–66.
22
Engin K, Leeper DB, Thistlethwaite AJ, Tupchong L, McFarlane JD. Tumor extracellular pH as a prognostic factor in thermoradiotherapy.
Int J Radiat Oncol Biol Phys
1994
;
29
:
125
–32.
23
van den Berg AP, Wike-Hooley JL, Broekmeyer-Reurink MP, van der Zee J, Reinhold HS. The relationship between the unmodified initial tissue pH of human tumours and the response to combined radiotherapy and local hyperthermia treatment.
Eur J Cancer Clin Oncol
1989
;
25
:
73
–8.
24
van den Berg AP, Wike-Hooley JL, van den Berg-Blok AE, van der Zee J, Reinhold HS. Tumour pH in human mammary carcinoma.
Eur J Cancer Clin Oncol
1982
;
18
:
457
–62.
25
Wike-Hooley JL, van den Berg AP, van der Zee J, Reinhold HS. Human tumour pH and its variation.
Eur J Cancer Clin Oncol
1985
;
21
:
785
–91.
26
Wike-Hooley JL, Van der Zee J, van Rhoon GC, Van den Berg AP, Reinhold HS. Human tumour pH changes following hyperthermia and radiation therapy.
Eur J Cancer Clin Oncol
1984
;
20
:
619
–23.
27
Walenta S, Schroeder T, Mueller-Klieser W. Lactate in solid malignant tumors: potential basis of a metabolic classification in clinical oncology.
Curr Med Chem
2004
;
11
:
2195
–204.
28
Walenta S, Wetterling M, Lehrke M, et al. High lactate levels predict likelihood of metastases, tumor recurrence, and restricted patient survival in human cervical cancers.
Cancer Res
2000
;
60
:
916
–21.
29
Brizel DM, Schroeder T, Scher RL, et al. Elevated tumor lactate concentrations predict for an increased risk of metastases in head-and-neck cancer.
Int J Radiat Oncol Biol Phys
2001
;
51
:
349
–53.
30
Kuznetsov YE, Caramanos Z, Antel SB, et al. Proton magnetic resonance spectroscopic imaging can predict length of survival in patients with supratentorial gliomas [discussion 74–6].
Neurosurgery
2003
;
53
:
565
–74.
31
Thrall DE, LaRue SM, Yu D, et al. Thermal dose is related to duration of local control in canine sarcomas treated with thermoradiotherapy.
Clin Cancer Res
2005
;
11
:
5206
–14.
32
Thrall DE, Rosner GL, Azuma C, et al. Using units of CEM 43 degrees C T90, local hyperthermia thermal dose can be delivered as prescribed.
Int J Hyperthermia
2000
;
16
:
415
–28.
33
Martinez-Zaguilan R, Seftor EA, Seftor RE, Chu YW, Gillies RJ, Hendrix MJ. Acidic pH enhances the invasive behavior of human melanoma cells.
Clin Exp Metastasis
1996
;
14
:
176
–86.
34
Kato Y, Lambert CA, Colige AC, et al. Acidic extracellular pH induces matrix metalloproteinase-9 expression in mouse metastatic melanoma cells through the phospholipase D-mitogen-activated protein kinase signaling.
J Biol Chem
2005
;
280
:
10938
–44.
35
Kato Y, Nakayama Y, Umeda M, Miyazaki K. Induction of 103-kDa gelatinase/type IV collagenase by acidic culture conditions in mouse metastatic melanoma cell lines.
J Biol Chem
1992
;
267
:
11424
–30.
36
Smallbone K, Gavaghan DJ, Gatenby RA, Maini PK. The role of acidity in solid tumour growth and invasion.
J Theor Biol
2005
;
235
:
476
–84.
37
Webb SD, Sherratt JA, Fish RG. Modelling tumour acidity and invasion [discussion 81–5].
Novartis Found Symp
2001
;
240
:
169
–81.
38
Vaupel PW, Frinak S, Bicher HI. Heterogeneous oxygen partial pressure and pH distribution in C3H mouse mammary adenocarcinoma.
Cancer Res
1981
;
41
:
2008
–13.
39
Helmlinger G, Yuan F, Dellian M, Jain RK. Interstitial pH and pO2 gradients in solid tumors in vivo: high-resolution measurements reveal a lack of correlation.
Nat Med
1997
;
3
:
177
–82.
40
Brizel DM, Dodge RK, Clough RW, Dewhirst MW. Oxygenation of head and neck cancer: changes during radiotherapy and impact on treatment outcome.
Radiother Oncol
1999
;
53
:
113
–7.
41
Rhee JG, Kim TH, Levitt SH, Song CW. Changes in acidity of mouse tumor by hyperthermia.
Int J Radiat Oncol Biol Phys
1984
;
10
:
393
–9.
42
Jahde E, Rajewsky MF, Baumgartl H. pH distributions in transplanted neural tumors and normal tissues of BDIX rats as measured with pH microelectrodes.
Cancer Res
1982
;
42
:
1498
–504.
43
Vaupel P, Schaefer C, Okunieff P. Intracellular acidosis in murine fibrosarcomas coincides with ATP depletion, hypoxia, and high levels of lactate and total Pi.
NMR Biomed
1994
;
7
:
128
–36.
44
Prescott DM, Charles HC, Poulson JM, et al. The relationship between intracellular and extracellular pH in spontaneous canine tumors.
Clin Cancer Res
2000
;
6
:
2501
–5.
45
Prescott DM, Charles HC, Sostman HD, et al. Therapy monitoring in human and canine soft tissue sarcomas using magnetic resonance imaging and spectroscopy.
Int J Radiat Oncol Biol Phys
1994
;
28
:
415
–23.
46
Evanochko WT, Sakai TT, Ng TC, et al. NMR study of in vivo RIF-1 tumors. Analysis of perchloric acid extracts and identification of 1H, 31P and 13C resonances.
Biochim Biophys Acta
1984
;
805
:
104
–16.
47
de Certaines JD, Larsen VA, Podo F, Carpinelli G, Briot O, Henriksen O. In vivo 31P MRS of experimental tumours.
NMR Biomed
1993
;
6
:
345
–65.
48
Garcia-Martin ML, Martinez GV, Raghunand N, Sherry AD, Zhang S, Gillies RJ. High resolution pH(e) imaging of rat glioma using pH-dependent relaxivity.
Magn Reson Med
2006
;
55
:
309
–15.
49
Gillies RJ, Raghunand N, Garcia-Martin ML, Gatenby RA. pH imaging. A review of pH measurement methods and applications in cancers.
IEEE Eng Med Biol Mag
2004
;
23
:
57
–64.
50
Ojugo AS, McSheehy PM, McIntyre DJ, et al. Measurement of the extracellular pH of solid tumours in mice by magnetic resonance spectroscopy: a comparison of exogenous (19)F and (31)P probes.
NMR Biomed
1999
;
12
:
495
–504.
51
Gil S, Zaderenzo P, Cruz F, Cerdan S, Ballesteros P. Imidazol-1-ylalkanoic acids as extrinsic 1H NMR probes for the determination of intracellular pH, extracellular pH and cell volume.
Bioorg Med Chem
1994
;
2
:
305
–14.