Twenty consecutive patients with squamous cell carcinoma of the uterine cervix who received definitive RT, with or without chemotherapy, at Samsung Medical Center between March 2005 and November 2006 were evaluated (Table 1). When pelvic lymph nodes exceeded 1.5 cm in diameter and/or there was significant F18-fluorodeoxyglucose uptake by positron emission tomography (PET), the lymph node was considered positive for cancer involvement. Pre-treatment PET was performed in 12 patients. Among them, seven patients had significant abnormal uptake in pelvic lymph nodes. The institutional review board approved this study and all patients voluntarily participated in this study by signing the informed consent.

All patients received a combination of external beam radiotherapy (EBRT) and high-dose-rate (HDR) intracavitary irradiation (ICR) by a remote after loading system with Ir-192. EBRT was delivered to the whole pelvis with 15-MV photon beams at a daily dose of 1.8 Gy, five times per week to a total dose of 50.4 Gy. The patients were irradiated with a 4-field box technique in order to spare the small bowel anterior to the iliac nodes. Extended field irradiation including the para-aortic area, with a total dose of 45 Gy, was adminstered for one patient with positive common iliac lymph nodes. HDR ICR was initiated after an EBRT dose of 41.4 Gy to 45 Gy with midline blocking. ICR was delivered twice a week in six fractions with a fractional dose of 4 Gy at point A. The median overall treatment time was 53 days (range, 46 to 61 days). Three patients with FIGO stage IB or IIA disease who had small tumors less than 3 cm did not receive concurrent chemotherapy. The remaining seventeen patients received concurrent chemoradiotherapy. Eleven patients received six cycles of weekly cisplatin (30 mg/m²) during RT. The remaining six patients were treated with a 5-fluorouracil (1,000 mg/m²) plus cisplatin (60 mg/m²) regimen every 3-4 weeks for 2 or 3 cycles during RT, and additional chemotherapy were administered after RT for 1 or 3 cycles. The selection of the chemotherapeutic regimen was individualized according to local tumor extent, pelvic lymph node involvement, and patient's general condition. During treatment, assessment of acute toxicity was performed every week using the Radiation Therapy Oncology Group (RTOG) criteria.

Serial MRI examinations were performed at three time points: at the start of the RT (pre-RT), at the fourth week of RT (mid-RT), and 1 month after completion of RT (post-RT). MRI was performed with a 1.5-T MR system (Signa; GE Medical Systems, Milwaukee, WI, USA) using a phased array or body coil. The pelvis was imaged with an axial fast spinecho T2-weighted sequence with an echo train length of 20 to 25, a 5-mm-section thickness, and a 2-mm intersection gap. The matrix size was 512×256 and the field of view was 28 to 30 cm. Three-dimensional tumor volumes were calculated by a radiation oncologist without knowledge of the FIGO stage and tumor marker status. On T2-weighted images, the area of the tumor in each slice was outlined and the tumor volume was calculated by summation of all the areas of tumor and multiplied by the slice interval. Tumor volume responses were calculated by dividing the tumor volume at a given point by the pre-RT tumor volume (V1),

where V2 was the mid-RT tumor volume, V3 was the post-RT tumor volume; V2R and V3R were the percentage of residual tumor volume at each time.

Serial punch biopsies were performed at pre-RT and mid-RT with intent to obtain tumor tissues adequately. After verifying the presence of the tumor on the hematoxylin and eosin stained slides, immunohistochemical staining was performed for COX-2 and EGFR. Formalin-fixed, paraffin-embedded 5 µm-thick sections were prepared from the biopsy specimens. The slides were deparaffinized in xylene and endogenous peroxidase was blocked by incubating the slides in 0.3% H₂O₂ for 15 minutes. Antigen retrieval was performed using 0.01 M citrate buffer for COX-2 and proteinase K for EGFR, respectively. Incubation of the sections in a solution of 4% bovine serum and phosphate buffered saline dilution for 30 minutes was performed to block nonspecific binding. Then, the specimen was incubated with the primary antibody. The COX-2 antibody was purchased from Cayman (Ann Arbor, MI, USA) and the EGFR antibody from Dako (Carpinteria, CA, USA). The sections were incubated with anti-mouse secondary antibodies of the ChemaMate TM detection kit from Dako. Deaminobenzidine was used as the chromogen, and hematoxylin was used as the nuclear counter stain.

Two experienced pathologists, without knowledge of the clinical information, examined the sections under the light microscope and evaluated the distribution of the immunoreactivity. The intensity of staining was classified as negative indicating no expression, weak indicating minimal staining, moderate indicating moderate staining, and strong representing heavy staining. If the distribution of moderate to strong immunoreactivity accounted for more than 10%, the specimen was classified as positive. If there was an available mid-RT immunohistochemical stained slide, the distribution of moderate to strong immunoreactivity was compared with that of the pre-RT specimen. The alteration of the distribution of immunoreactivity was analyzed.

The relationship between the distribution of immunoreactivity for COX-2 and EGFR was assessed by Pearson's correlation. Pearson's correlation was also applied for assessing the relationship between V1 and V2R. Correlations between tumor volume parameters (V1, V2R, V3R) and biological markers (COX-2, EGFR) were examined by Fisher's exact test. For identifying factors associated with the mid-RT tumor volume response (V2R), a multiple logistic regression analysis was employed including V1, FIGO stage, and positivity for both biological markers. The SAS ver. 9.0 (SAS Institute Inc., Cary, NC, USA) was used for statistical analysis. A p-value ≤ 0.05 was considered statistically significant.

The median tumor volume at pre-RT was 20.9 cm³ (range, 1.7 to 132.3 cm³), which regressed to 9.3 cm³ (range, 0.4 to 73.5 cm³) at mid-RT. The percentage of residual tumor volume at mid-RT (V2R) ranged from 0.136 to 0.983 (median 0.396). The relationship between V1 and V2R was not statistically significant (Pearson correlation coefficient=.0.1376, p=0.5629) (Fig. 1). At post-RT, 15 (75%) patients showed complete regression of tumor on MRI. The remaining five (25%) patients had residual tumor that ranged from 1.7 cm³ to 6.6 cm³.

For the pre-RT evaluation, fourteen (70%) and eleven (55%) patients showed positive immunoreactivity for COX-2 and EGFR, respectively (Fig. 2). Among them, eight patients (40%) showed positive immunoreactivity for both biological markers. The relationship between the distribution of immunoreactivity for the biological markers was not statistically significant (Pearson correlation coefficient=0.2985, p=0.2011).

For the mid-RT evaluation, there were only twelve available immunohistochemical stained slides for COX-2 and EGFR (60%). The reasons that not all patients had slides available for review included tumor necrosis by radiation and/or sampling errors. Among the twelve available slides at mid-RT, eight (67%) cases for COX-2 and six (50%) cases for EGFR showed an interval increase in the distribution of immunoreactivity, compared with immunoreactivity of the pre-RT evaluation.

Among the seven patients whose V2R was greater than 0.5, all seven (100%) showed positive immunoreactivity for COX-2, and five (71.4%) for EGFR at the pre-RT evaluation (Table 2). The relationship between the V2R and the pre-RT COX-2 expression was marginally significant (p=0.0515). There was no significant relationship between the other volume factors (V1, V3R) and the biological markers. Among the five patients whose V3R was greater than 0, three (60%) and two (40%) showed positive immunoreactivity for COX-2 and EGFR, respectively.

Among the eight patients who showed positive immunoreactivity for both biological markers, a V2R greater than 0.5 was observed in five patients (63%) compared with the rest (17%). However, the difference was only marginally significant (p=0.0623). The logistic regression analysis showed that positive immunoreactivity for both biological markers was associated with the V2R (p=0.0782), which was also marginally significant (Table 3).

Among the eight patients who showed an interval increase in the distribution of immunoreactivity for COX-2 at mid-RT compared with immunoreactivity for COX-2 at pre-RT, six (75%) patients had V2R greater than 0.5 compared with the patients who showed no change or a decrease in the distribution of COX-2 expression (25%) (Table 4). However, the finding was not statistically significant (p=0.2222). Two out of the six patients showed residual tumor at the post-RT evaluation. There were no significant findings for the interval change in the distribution of immunoreactivity for EGFR (data not shown).

There was no incidence of RTOG grade 3 or higher radiation related toxicity. Six patients experienced grade 2 toxicity, five experienced lower gastrointestinal and one genitourinary toxicity. Among the six patients, four patients showed positive immunoreactivity for COX-2 at the pre-RT evaluation and the remaining two patients did not.