Abstract
Purpose: To investigate tumor interstitial fluid pressure as a prognostic factor for recurrence-free survival in patients with cervical cancer following radiation therapy.
Experimental Design: Tumor interstitial fluid pressure was measured in 55 cervical cancer patients who received radiation therapy between August 1998 and September 2002. Interstitial fluid pressure measurements were made before radiation therapy (pre–radiation therapy interstitial fluid pressure) and after a median of 28.8 Gy in 16 fractions (range, 25.2-30.6 Gy in 14-17 fractions) of radiation therapy (mid–radiation therapy interstitial fluid pressure), using a modified wick-in-needle technique. Median follow-up was 74 months (range, 2-118 months). The Kaplan-Meier method with the log-rank test and Cox's proportional hazard model were used in univariate and multivariate analyses, respectively, of prognostic factors for recurrence-free survival.
Results: Median pre–radiation therapy and mid–radiation therapy interstitial fluid pressure were 29.0 mm Hg (range, 4.0-93.9 mm Hg) and 20.0 mm Hg (range, −1.2 to 29.6 mm Hg), respectively (P = 0.001). Pre–radiation therapy interstitial fluid pressure was significantly higher in adenocarcinomas than squamous cell carcinomas (P = 0.028). Significant reduction of interstitial fluid pressure was noted only in patients with complete responses (P = 0.002), and mid–radiation therapy interstitial fluid pressure was significantly lower in patients with complete responses (P = 0.036). In the multivariate analysis including interstitial fluid pressures and clinical variables, pre–radiation therapy interstitial fluid pressure was an independent prognostic factor for local and distant recurrence-free survival (P = 0.001 and 0.027, respectively).
Conclusions: Mid–radiation therapy interstitial fluid pressure measurement may be useful in predicting radiation therapy responses, and pre–radiation therapy interstitial fluid pressure was a significant prognostic factor for local and distant relapse-free survival in patients with cervical cancer after radiation therapy. (Clin Cancer Res 2009;15(19):6201–7)
- Interstitial fluid pressure
- Cervical cancer
- Radiation therapy
- Prognostic factor
- Survival
Translational Relevance
The purpose of this study was to investigate interstitial fluid pressure, measured before and during treatment, as a predictor of response and long-term treatment outcome in patients with cervical cancer following radiation therapy. We measured interstitial fluid pressure before treatment and after a median of 28.8 Gy in 16 fractions of radiation therapy in 55 patients with cervical cancer after obtaining informed consent. The Kaplan-Meier method with the log-rank test and Cox's proportional hazard model were used in univariate and multivariate analyses, respectively, of prognostic factors for recurrence-free survival. The results of this study will be useful in selecting cervical cancer patients predicted to have less favorable radiation responses or to be at high risk for disease recurrence after radiation therapy.
Interstitial fluid pressure in normal tissue is close to atmospherical pressure, whereas interstitial fluid pressure in solid tumors is increased by 10 to 100 mm Hg (1–5). This increased tumor interstitial fluid pressure is thought to be the result of unregulated angiogenesis, a lack of functional lymphatics, and abnormal extracellular matrix. The vessels produced in tumor angiogenesis are hyperpermeable, and the resultant net efflux of fluid into the interstitium, with impaired lymphatic drainage, increases interstitial fluid volume and distends the elastic extracellular matrix. In addition, tumor proliferates in confined space and tumor interstitium consists of dense collagen fibers and increased inflammatory components, such as fibroblasts and macrophages, contributing to the increased interstitial fluid pressure (6–10).
Interstitial fluid pressure can be measured easily in accessible human tumors using a simple needle probe. Interstitial fluid pressure is thought to reflect the global physiology of a tumor, which is largely unrelated to perfusion or oxygen status (11, 12). Thus, tumor interstitial fluid pressure may be of prognostic use in the treatment of patients with cancer. To date, interstitial fluid pressure measurement and long-term follow-up after radiation therapy has only been reported at one institution; pretreatment interstitial fluid pressure was shown to be able to predict the radiation therapy response (13) and was an important prognostic factor for local and distant disease progression in patients with cervical cancer receiving radiation therapy (5, 14). However, to our knowledge, no reported study has investigated interstitial fluid pressures measured during treatment and pretreatment to assess their prognostic significance.
We measured tumor interstitial fluid pressure before and during radiation therapy in patients with cervical cancer managed with definitive radiation therapy. The purpose of this study was to investigate whether interstitial fluid pressure can be a prognostic factor for long-term treatment outcomes in patients with cervical cancer following definitive radiation therapy. In addition, the correlation of interstitial fluid pressure with patient characteristics was analyzed, as well as the potential of interstitial fluid pressure as an early predictor of treatment response.
Materials and Methods
Patients
This was a single-institution prospective clinical study. Tumor interstitial fluid pressure was measured in 55 patients with uterine cervix cancer who received definitive radiation therapy at Chungnam National University Hospital between August 1998 and September 2002. The study was done in accordance with the guidelines of our Institutional Review Board. Written informed consent was obtained from each patient before interstitial fluid pressure measurements. We followed these patients for a period sufficient to investigate the prognostic significance of interstitial fluid pressure, and this is the first report on the long-term results of the study.
The patients with cervical tumors that could be visualized vaginally with a minimum tumor thickness of 8 mm were chosen for interstitial fluid pressure measurement because of the need for adequate placement of the measuring needle tip. The characteristics of the patients included in this study are summarized in Table 1. The median age was 56 y (range, 32-77 y). The histology was squamous cell carcinoma in 49 (89%), adenocarcinoma in five (9%), and neuroendocrine cell carcinoma in one (2%) patient. International Federation of Gynecology and Obstetrics (FIGO) stage was Ib in 12 (22%), IIa in five (9%), IIb in 33 (60%), IIIa in one (2%), and IIIb in four (7%) patients. Pelvic and para-aortic lymph node status was evaluated with pelvic magnetic resonance imaging in 47 (85%), abdominopelvic computed tomography in six (11%), and both modalities in two (4%) patients; it was positive in 16 (29%) and two (4%) patients, respectively. The radiological criterion for positive lymph node metastasis was size ≥1 cm in the shortest axis diameter. The pre–radiation therapy serum hemoglobin level was 5.0 to 14.3 g/dL (median, 12.4 g/dL).
Patient characteristics (N = 55)
Measurement of interstitial fluid pressure and tumor size
Measurement of interstitial fluid pressure and tumor size was done in two periods, before starting radiation therapy (pre–radiation therapy interstitial fluid pressure) and after 25.2 to 30.6 Gy in 14 to 17 fractions of radiation therapy (mid–radiation therapy interstitial fluid pressure). Interstitial fluid pressure was measured using a modified wick-in-needle technique. The patients were in the lithotomic position, with the vertical position of the cervical tumor at approximately heart level. A 25-gauge needle with 3-mm side hole located 5 mm from the needle tip was filled with a nylon surgical suture filament (6-0 Ethilon). This filament was exposed to the tissue through the side hole. The needle was connected to a pressure transducer (Model II-G4113-01; Gould Electronics) through a polyethylene tube filled with heparinized isotonic saline (70 units/mL). The pressure signal was amplified and stored in a recording system (Model BD111; Kipp and Zonen). The needle and polyethylene tube were sterilized before each measurement. Calibration was carried out before and after each measurement by making pressures of 0, 16.5, and 23 mm Hg at heights of 0, 15, and 30 cm, respectively. A needle was manually advanced into the central region of the tumor to a depth such that at least the entire side hole of the needle was within the tumor and was left in place without external fixation. Fluid communication during measurement was checked by compressing and decompressing the polyethylene tube with a clamp. If the pressure values during 5 to 10 min were within ±3 mm Hg, the measurement was considered valid. The measurement was repeated in the central region of the tumor within 5 min, and interstitial fluid pressure was determined as the mean of the two values. Interstitial fluid pressure measurements were well tolerated without using anesthesia.
Pre–radiation therapy tumor size was determined as the largest axis diameter measured by clinical examination and abdominopelvic computed tomography or pelvic magnetic resonance imaging. Mid–radiation therapy tumor size was measured only by clinical examination.
Radiation therapy
Radiation therapy consisted of external-beam radiation therapy and high-dose-rate intracavitary radiation therapy using a remote afterloading technique with an 192Ir source. External-beam radiation therapy was administered using 6 or 10 MV X-rays, 5 days a week and 1.8 Gy per fraction. The four-field box technique or opposed anteroposterior and posteroanterior fields were used for the treatment of the pelvis or the pelvis and para-aortic region. The pelvic field included the primary tumor and lymph nodes of the internal, external, and lower common iliac chain. After 36.0 to 45.0 Gy (median, 39.6 Gy) of external-beam radiation therapy, a midline shield block was added and high-dose-rate intracavitary radiation therapy was given in six to seven fractions, twice a week, concurrent with parametrial boost. Intracavitary radiation therapy dose at point A was 19.7 to 41.0 Gy (median, 31.1 Gy). The total cervical dose, including the pelvic dose, before parametrial boost and intracavitary radiation therapy dose was 57.4 to 86.0 Gy (median, 70.7 Gy). The total external-beam radiation therapy dose, including the pelvic dose and parametrial boost, was 45.0 to 61.2 Gy (median, 54.0 Gy). For three patients in whom intracavitary radiation therapy was not done because of poor geometry (unfavorable tumor anatomy), 50.4 Gy with pelvic fields and additional 19.8 Gy with reduced fields was given (total, 70.2 Gy). Chemotherapy was not used.
Treatment response was evaluated by pelvic examination and abdominopelvic computed tomography or pelvic magnetic resonance imaging. The criteria for a response followed the guidelines developed by the WHO (15). Patients were followed at 3-mo intervals for the first year, 4-mo intervals for the second and third years, and 6-mo intervals thereafter. At each visit, a physical examination, including pelvic examination, laboratory tests, including tumor markers, and abdominopelvic computed tomography or pelvic magnetic resonance imaging, was conducted.
Statistical analysis
The paired samples t test was used for the mean comparison between pre–radiation therapy and mid–radiation therapy values of interstitial fluid pressure or tumor size. The relationship between pre–radiation therapy interstitial fluid pressure and patient characteristics was evaluated using Pearson's r, Student's t test, and the χ2 or Fisher's exact test, as appropriate. The box plot was used to show median, quartiles, and extreme values of interstitial fluid pressure. The variation of two interstitial fluid pressures in each tumor at each period of measurement was expressed as the numerical difference of two values divided by lower value (percentage). This variation was analyzed for the relationship with the magnitude of mean of two interstitial fluid pressures and tumor size using Pearson's r.
For the analysis of predictors of treatment response, logistic regression and the χ2 or Fisher's exact test was used as appropriate. Events in recurrence-free survival were defined as progression or recurrence of disease, or patient death due to any cause. Local recurrence was defined as progression or recurrence in the pelvic region, and distant metastasis was defined as recurrence outside the pelvic region. Survival was estimated from the start of radiation therapy using the Kaplan-Meier method and compared by the log-rank test. For multivariate analysis of prognostic factor for survival, Cox's proportional hazard model with backward stepwise selection of covariates was used, with a P of <0.1 adopted as the threshold for covariate inclusion. Two multivariate analyses were done, including all potential clinical parameters with and without adding interstitial fluid pressures. Age, hemoglobin, and tumor size were included as continuous variables and the others as binary parameters. Adjusted survival curves were constructed using multivariate Cox's proportional hazard model with average covariate method (16). A P of <0.05 was deemed to indicate a statistically significant difference.
Results
Relationship between interstitial fluid pressure and patient characteristics
The median pre–radiation therapy interstitial fluid pressure and tumor size were 29.0 mm Hg (range, 4.0-93.9 mm Hg) and 3.0 cm (range, 1.6-6.0 cm), respectively. The mid–radiation therapy measurements of interstitial fluid pressure and tumor size were done after a median of 28.8 Gy in 16 radiation therapy fractions. The mid–radiation therapy interstitial fluid pressure and tumor size were not measured in 10 and 12 patients, respectively, because of significant tumor size reduction, severe bleeding, or refusal of further radiation therapy after 9.0 Gy in five fractions in one patient. The median mid–radiation therapy interstitial fluid pressure and tumor size were 20.0 mm Hg (range, −1.2 to 79.6 mm Hg) and 2.5 cm (range, 0.5-4.5 cm), respectively. Median residual tumor size was 75% (range, 33-133%), and mean differences of interstitial fluid pressure and tumor size between the measurements were statistically significant (interstitial fluid pressure, P = 0.001; tumor size, P < 0.001; Fig. 1). Pre–radiation therapy interstitial fluid pressure was significantly higher in adenocarcinomas than squamous cell carcinomas (49.2 ± 18.8 and 29.9 ± 18.2 mm Hg, respectively; P = 0.028). No relationship was observed between pre–radiation therapy interstitial fluid pressure and pre–radiation therapy patient characteristics, such as age, stage, hemoglobin, pelvic lymph node status, or tumor size. No relationship was seen between mid–radiation therapy interstitial fluid pressure and mid–radiation therapy tumor size, or interstitial fluid pressure change and tumor size change. Pre–radiation therapy tumor size was significantly larger in the patients with clinically positive pelvic lymph nodes than those with negative pelvic lymph nodes (3.8 ± 1.1 and 3.1 ± 0.9 cm, respectively; P = 0.013). Comparing two interstitial fluid pressures in each tumor, the pressures varied by 3% to 63% (median, 19%) and 10% to 87% (median, 32%) at pre–radiation therapy and mid–radiation therapy, respectively. No significant correlation was observed between this variation and mean interstitial fluid pressure or tumor size in each tumor, either at pre–radiation therapy or mid–radiation therapy.
Box plot showing significant reduction of interstitial fluid pressure (millimeters of mercury) after median 28.8 Gy in 16 fractions of radiation therapy (P = 0.001). Top and bottom of each box, the interquartile range; the whiskers mark the highest and lowest within 1.5 times the interquartile range. Outliers (O) are values between 1.5 and 3.0 times the interquartile range from the edge of the box, and extreme (*) values are more than thrice the interquartile range from the edge of box.
Interstitial fluid pressure as a predictor of radiation therapy response
Two patients did not finish the scheduled radiation therapy and were excluded from the analysis of predictors of treatment response and survival. Treatment response evaluated at 1 month after radiation therapy had finished was a complete response in 38 (72%), partial response in 10 (19%), and progressive disease in one (2%) patient, a para-aortic lymph node metastasis. Four (7%) patients were not assessable: two patients were lost to follow-up, one patient visited the clinic at 4 months, and one patient died after radiation therapy in an accident. Treatment response evaluated at 3 to 4 months after radiation therapy had finished was a complete response in 42, partial response in four, and progressive disease in two patients. Significant reduction of interstitial fluid pressure occurred only in patients with a complete response at 1 month (30.1 ± 17.4 and 20.3 ± 10.4 mm Hg for pre–radiation therapy and mid–radiation therapy interstitial fluid pressure, respectively; P = 0.002). Reduction of interstitial fluid pressure in patients with a partial response at 1 month was not significant (33.2 ± 22.0 and 31.3 ± 20.9 mm Hg for pre–radiation therapy and mid–radiation therapy interstitial fluid pressure, respectively; P = 0.732). Mid–radiation therapy interstitial fluid pressure was significantly different between patients with complete and partial responses at 1 month (P = 0.036). No other clinical factor, including tumor size, that could predict treatment response was found.
Interstitial fluid pressure as a prognostic factor of recurrence-free survival
The median follow-up was 74 months (range, 2-118 months). Local progression or recurrence occurred in three (6%), distant metastasis in nine (17%), and concurrent local and distant recurrence in one (2%) patient. The most common site of distant metastasis was the para-aortic lymph node (four patients); other sites included the lung, liver, and supraclavicular lymph node. Median time to disease progression or recurrence was 15 months (range, 3-92 months). At last follow-up, 19 (36%) patients had died. The 5-year overall survival rate was 66%. Median recurrence-free survival and 5-year recurrence-free survival rates were 108 months and 63%, respectively. In a univariate analysis of prognostic factors, histology and radiation therapy response at 1 month were significant prognostic factors for local and distant recurrence-free survival (Table 2). Pre–radiation therapy interstitial fluid pressure, either as binary or continuous variable, was not statistically significant in the univariate prognostic factor analysis. In a multivariate analysis including only clinical parameters, significant prognostic factors for local or distant recurrence-free survival were age and radiation therapy response at 1 month. When interstitial fluid pressures were added to this multivariate analysis, pre–radiation therapy interstitial fluid pressure was found to be an independent prognostic factor for local and distant recurrence-free survival (Table 3). Pre–radiation therapy interstitial fluid pressure was a significant prognostic factor for recurrence-free survival after adjusting for the influence of confounding covariates in the multivariate analysis (Figs. 2 and 3).
Univariate analysis of prognostic factors for recurrence-free survival
Multivariate analysis of prognostic factors for recurrence-free survival
Adjusted survival curves comparing local recurrence–free survival according to pretreatment interstitial fluid pressure (<29 versus ≥29 mm Hg) after correcting for the influence of confounding covariates using multivariate Cox's proportional hazard model (P = 0.001).
Adjusted survival curves comparing distant metastasis–free survival according to pretreatment interstitial fluid pressure (<29 versus ≥29 mm Hg) after correcting for the influence of confounding covariates using multivariate Cox's proportional hazard model (P = 0.027).
Discussion
To our knowledge, prognostic implications of interstitial fluid pressure in cancer patients after long-term follow up have been reported by only one institution (5, 13, 14). Interstitial fluid pressure and oxygenation were measured before radiation therapy in 107 cervical cancer patients under general anesthesia, and the median interstitial fluid pressure was 19.0 mm Hg (range, −2.8 to 48.0 mm Hg). They reported that the pretreatment interstitial fluid pressure was a significant prognostic factor, independent of standard clinical prognostic factors and tumor hypoxia, for local and distant relapse-free survival (14). We measured tumor interstitial fluid pressure in 55 patients with cervical cancer, and the median pre–radiation therapy interstitial fluid pressure was 29.0 mm Hg (range, 4.0-93.9 mm Hg). Excluding three outliers by the box plot, the values between 1.5 and 3.0 times interquartile range from upper edge of interquartile range, pre–radiation therapy interstitial fluid pressure ranged from 4.0 to 65.0 mm Hg. Anesthesia was not used during interstitial fluid pressure measurement in this study, but it is thought that anesthesia has no major effect on interstitial fluid pressure (13). We measured interstitial fluid pressure twice only in central region of tumor, but the authors of institution aforementioned did multiple site measurements (typically, 4-5) around tumor circumference for consideration of regional heterogeneity (13). However, there are reports that interstitial fluid pressure is relatively uniform throughout a tumor growing as a single nodule and that it drops precipitously at the tumor-normal tissue interface (7, 17, 18). Despite some differences in methods of interstitial fluid pressure measurement, the result of this study was similar; pre–radiation therapy interstitial fluid pressure was a prognostic factor for local and distant relapse-free survival in cervical cancer patients undergoing radiation therapy.
A few studies have been published about interstitial fluid pressure measurements pre–radiation therapy and/or during radiation therapy for the purpose of prediction of radiation therapy response. In the study of Roh et al. (2), the first interstitial fluid pressure measurement was made at 7.2 Gy because of bleeding control with radiation therapy and the second one after 30 Gy of radiation therapy. Among seven cervical cancer patients with both measurements, four patients with interstitial fluid pressure reduction fully responded, whereas two of the three patients with interstitial fluid pressure increases had moderate responses evaluated 4 to 6 weeks after radiation therapy had ended. In the study by Milosevic et al. (13), the treatment response evaluated at 3 months after radiation therapy was predictable by interstitial fluid pressure. In our study, we evaluated radiation therapy responses at 1 and 3 to 4 months after radiation therapy was completed, and the number of complete responders increased by four between the periods. Significant interstitial fluid pressure decreases occurred only in patients with complete responses, and mid–radiation therapy interstitial fluid pressure was significantly lower in patients with complete response than in those with partial responses at 1 month after radiation therapy. In addition, the radiation therapy response at 1 month after radiation therapy was shown to be an independent prognostic factor for local and distant relapse-free survival. Thus, interstitial fluid pressure measurements after early treatment, in addition to pretreatment interstitial fluid pressure, may help to predict radiation therapy responses and long-term prognosis and, thus, to determine which patients likely need modification of treatment.
Five (9%) patients with cervix adenocarcinoma were included in this study. Although the number was small, these tumors had higher pre–radiation therapy interstitial fluid pressure compared with squamous cell carcinomas. The two histologic groups had no significant difference in age, stage, pelvic lymph node status, or tumor size. Several studies have reported that vascular endothelial growth factor (VEGF) secretion or expression and microvessel density was higher in adenocarcinomas than in squamous cell carcinomas in cervical cancer (19–21). VEGF, a potent factor in angiogenesis, may be related to the high interstitial fluid pressure in adenocarcinomas because increased interstitial fluid pressure is largely due to unregulated angiogenesis. Several previous studies have reported lower radiation responses and less favorable prognoses in patients with adenocarcinomas than in those with squamous cell carcinomas in cervical cancer (22, 23). Further studies investigating the relationship of interstitial fluid pressure with histologic types will be valuable in clarifying the pathophysiology of high tumor interstitial fluid pressure and developing targeting agents to overcome the radioresistance of adenocarcinomas.
The most important mechanism postulated to underlie the relationship between interstitial fluid pressure and radiation therapy has been that high interstitial fluid pressure may provide a relative indication of tumor angiogenic activity and vascular persistence as a cause of radioresistance (12, 14). Blocking angiogenesis may seem to counteract a radiation response because of diminished blood perfusion and the resulting hypoxia. However, many laboratory and clinical studies have provided evidence showing an opposite effect. One hypothesis is that antiangiogenic treatment causes vasculature normalization or stabilization, resulting in improved delivery of drugs or radiosensitizing oxygen (24). Examples of antiangiogenic agents have included VEGF-receptor signaling inhibitor (25–27), anti-VEGF monoclonal antibody (28, 29), and agents directly targeting the vasculature (30–32). Combining these therapeutic implications of interstitial fluid pressure and the prognostic value of interstitial fluid pressure observed in this study, it is reasonable to suggest that patients with high pre–radiation therapy interstitial fluid pressure or low interstitial fluid pressure reduction at mid–radiation therapy will need intensified administration of these molecular targeted agents to enhance radioresponse and overall treatment outcomes.
This study was done before the introduction of the combined modality treatment of concurrent chemoradiotherapy, which is now the standard cytotoxic management for cervical cancer in our hospital. Several cooperative oncology groups have shown in randomized studies that combining cisplatin-based chemotherapy concurrently with radiation therapy improves outcomes for patients with cervical cancer, acting as a potential radiation sensitizer (33, 34). Combining the results of a preclinical study showing that lowering tumor interstitial fluid pressure augments the efficacy of chemotherapy by enhancing drug delivery (35), it might be theorized that the prognostic implication of interstitial fluid pressure can be shown (or augmented) when concurrent chemoradiation is used (compared with radiation therapy alone). In a recent clinical study, the combination of an anti-VEGF antibody, bevacizumab, with concurrent preoperative chemoradiation showed interstitial fluid pressure reduction and improvement of outcomes for rectal cancer (29). However, further studies will be needed to investigate the prognostic significance of interstitial fluid pressure in patients with cervical cancer receiving concurrent chemoradiotherapy.
We measured mid–radiation therapy interstitial fluid pressure relatively early after giving less than half of the total radiation dose and tried to measure interstitial fluid pressure in the central region of the residual lesion with weekly pelvic examinations during treatment. However, interstitial fluid pressure might have been measured in regressed nonmalignant tissues, considering the possibility of nonconcentric irregular shrinkage of cervical tumors responding to radiation therapy (36). To measure interstitial fluid pressure during treatment, additional magnetic resonance imaging of the pelvis before mid–radiation therapy interstitial fluid pressure measurement would be helpful.
In conclusion, the results of this study indicate that interstitial fluid pressure provides clinically relevant information about the tumor microenvironment and helps predict treatment outcome in patients with cervical cancer receiving radiation therapy. Further studies to clarify the detailed mechanism underlying the biological linkage between interstitial fluid pressure and radiation therapy outcome are warranted. The modulation of interstitial fluid pressure with radiation therapy may provide a valuable therapeutic strategy.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Footnotes
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.
- Received March 4, 2009.
- Revision received June 14, 2009.
- Accepted June 26, 2009.