Relationships between thermal dose parameters and the efficacy of definitive chemoradiotherapy plus regional hyperthermia in the treatment of locally advanced cervical cancer: data from a multicentre randomised clinical trial

Abstract

Purpose: To evaluate the contribution of the thermal dose parameters during regional hyperthermia (HT) treatment to the clinical outcomes in patients with cervical carcinoma (CC) who received chemoradiotherapy (CRT) plus HT.

Materials and methods: Data from a multicentre randomised clinical trial of concurrent CRT + HT vs. CRT alone were used to evaluate the efficacy and safety of this combination therapy in the CC patients. The intrarectal temperatures of patients undergoing HT were recorded. The complete thermal data of 47 (92%) of the 51 patients in the CRT + HT group were available for the thermal analysis. Thus, 47 patients who received CRT + HT were included in the present study.

Results: Among the patients who received CRT + HT, a higher CEM43T90 (≥1 min) value (a thermal dose parameter) was significantly associated with better local relapse-free survival in both univariate (p = 0.024) and multivariate (p = 0.0097) analyses. The disease-free survival of the patients with higher CEM43T90 (≥1 min) values tended to be better in comparison to patients with lower CEM43T90 (<1 min) value (p = 0.071). A complete response tended to be associated with the CEM43T90 (p = 0.056). Disease-free survival, local relapse-free survival and complete response rate for patients with higher CEM43T90 (≥1) were significantly better than those for patients with CRT alone (p = 0.036, p = 0.036 and p = 0.048).

Conclusions: Dose-effect relationships between thermal dose parameters and clinical outcomes were confirmed in the CC patients treated with a combination of CRT + HT. This study also confirmed that HT with lower CEM43T90 is insufficient to achieve a significant hyperthermic sensitisation to CRT.

Keywords:

• Cervical cancer
• hyperthermia
• chemoradiotherapyl thermal parameter
• radiosensitisation

Introduction

Concurrent cisplatin-based chemoradiotherapy (CRT) has been shown to improve overall survival (OS) in comparison to radiotherapy (RT) alone in patients with locally advanced cancer of the uterus cervix cervical carcinoma (CC). However, the prognosis of the disease is not satisfactory [1–3]. Thus, there is further room for improving the clinical outcomes of these patients. Hyperthermia (HT) is known to be cytotoxic to cancer cells and acts as a radiosensitizer and chemosensitizer [4–9]. The thermal sensitising effects with RT include the following reasons; radiation therapy-resistant tumour cells that are hypoxic, of low pH, nutritionally deprived, and in the S-phase are more sensitive to HT [10]. In 2000, a Dutch group reported the results of phase III study comparing RT alone and RT plus HT in the treatment of malignant pelvic tumours. In the study, significant improvements were observed in the OS and tumour complete response (CR) rates of patients with CC of the uterus who received RT + HT [11]. In the experimental setting, substantial radiosensitisation was observed at temperatures of ≥41 °C [12–19]. The results of a meta-analysis of the data from four phase III studies in which of HT was used to treat advanced and recurrent breast cancer revealed that the local tumour control rates were significantly better when tumours exhibited a higher intra-tumour temperature [20]. The thermal dose parameters were also correlated with the therapeutic outcomes in patients with several types of deep-seated tumour who were treated with RT + HT [21–24].

Recently, the results of a Japanese multicentre randomised clinical trial of concurrent CRT + HT vs. CRT alone, which was performed to evaluate the efficacy and safety of CRT + HT in patients with locally advanced CC, were reported [25]. In the randomised trial, CRT + HT resulted in a significantly improved CR rate in comparison to CRT; however, the survival-related outcomes such as local relapse-free survival (LRFS), disease-free survival (DFS) and OS did not differ to a statistically significant extent [25]. To our knowledge, there are no specific reports evaluating the dose-effect relationship of HT in CC patients treated with definitive CRT + HT, and only one report for preoperative CRT + HT has investigated this relationship in CC patients [26]. In this context, we hypothesised that there may be a positive dose-effect relationship between thermal dose parameters and clinical outcomes in CC patients treated with definitive CRT + HT. The purpose of this study was to assess the thermal dose parameters of regional HT and to explore the potential contribution of regional HT using the data from a Japanese multicentre randomised clinical trial of concurrent CRT + HT vs. CRT alone in patients with locally advanced CC.

Materials and methods

Patients

In the present study, we performed additional analyses of the thermal parameters in the data from a prospective, multicentre, randomised, parallel-group study. The study was conducted at five Japanese institutions by the Japanese Society of Hyperthermic Oncology (JASHO) group between September 2001 and March 2015. It aimed to evaluate the effectiveness of whole-pelvic HT using deep regional heating as well as standard CRT in the treatment of locally advanced CC. The study investigated the clinical response and survival of patients who were treated with cisplatin-based CRT vs. those who received CRT with HT (CRT + HT) [25]. The study protocol was approved by the institutional review boards at participating institutions. All participants provided written informed consent by signing a form that was approved by the institutional review boards of each of the centres. The inclusion and exclusion criteria in the randomised trial were previously reported [25]. The criteria are summarised in Table 1. In total, 101 patients with CC were enrolled in the trial; 50 patients were randomised into the CRT group and 51 were randomised into the CRT + HT group. The intrarectal temperatures of patients undergoing deep regional HT were recorded. The complete thermal data of 47 (92%) of the 51 patients in the CRT + HT group were available for the thermal analysis. Thus, 47 patients were included in the current study. The complete thermal data of the four remaining patients in the CRT + HT group were not available; they were therefore excluded from the study. Baseline characteristics and treatment of 47 patients in the CRT + HT group are listed in Table 2. Detailed characteristics and treatment of 50 patients in the CRT group have been described previously [25].

Table 1. The inclusion and exclusion criteria.

Criteria
Inclusion criteria
FIGO stages; IB bulky (≧4 cm diameter), IIA, IIB, IIIA, IIIB or IVA
ECOG performance status; 0–2
No history of RT, chemotherapy or surgery
Negative for para-aortic lymph nodes metastasis
No double cancer
No pacemaker
Adequate hematologic function
Adequate renal, hepatic and cardiac functions
Feasibility of cisplatin treatment
Feasibility of HT
Written informed consent
Exclusion criteria
Pregnancy or lactation
Distant metastasis
Serious, uncontrolled concurrent medical or neurologic conditions
Chronic heart failure or heart failure in the previous three months
Cerebral vascular disorder in the previous three months
Active infection

FIGO: international federation of gynaecology and obstetrics; ECOG: eastern cooperative oncology group.

Table 2. The patient characteristics and treatments.

Patient characteristics	CRT + HT (n = 47) (%)
Age (years), mean (range)	60.5 (30–87)
FIGO stage
IIA	0
IIB	11 (23)
IIIA	0
IIIB	32 (68)
IVA	4 (9)
Histology
Squamous cell ca.	40 (85)
Adenocarcinoma	3 (6)
Adenosquamous cell ca.	1 (2)
Small cell carcinoma	2 (4)
Undifferentiated ca.	1 (2)
Tumour size (cm), mean (range)	5.8 (3.2–14)
Treatment duration (days), mean (range)	51.1 (35–60)
External beam RT, mean (range)
Whole pelvis dose (Gy)	31.2 (20.0–45.0)
Central shield dose (Gy)	21.1 (10.0–21.6)
Brachytherapy
High-dose rate (Gy)	25.3 (20.0–30.0)
Total dose of cisplatin (mg/m^2), mean (range)	167.7 (30–200)
Total number of HT sessions
2	3
3	0
4	2
5	42
Follow-up duration (months), mean (range)	65.4 (2.5–156.3)

FIGO: international federation of gynaecology and obstetrics; CRT: chemoradiotherapy; HT: hyperthermia.

Chemoradiotherapy

Chemoradiotherapy was performed according to a previously described protocol [25]. In brief, whole-pelvis RT using 6- or 10-MV high-energy linear accelerators was performed, and an additional dose was given to the parametrium with central shielding. RT was delivered with anteroposterior and posteroanterior parallel-opposed ports or four-field box in fractions of 1.8–2 Gy daily, five days per week. Additionally, patients also received high-dose rate intracavitary brachytherapy, which was administered at point A (5–6 Gy per session once per week). The patients were also treated weekly with cisplatin (30–40 mg/m² for 3–5 cycles).

Hyperthermia and thermal parameters

In all cases, HT treatment consisted of 8 MHz radiofrequency-capacitive regional HT using a Thermotron RF-8 (Yamamoto Vinita Co., Osaka, Japan). The physical features of this instrument and the thermal distribution in a phantom model and the human body have been reported previously [27–29]. The median duration of heating in one heating session was 50 min (30–50). The number of heating sessions ranged from 2 to 5 (median, 5) (Table 2). Regional whole-pelvis HT was administered on a weekly basis concurrent with cisplatin-CRT, but not with brachytherapy. Successive treatments were administered in the following order: cisplatin, RT and HT. External RT was performed less than 1 h after the administration of cisplatin. Deep regional HT was applied within 30 min of external RT during an external beam RT course. The three therapeutic modalities (cisplatin, RT and HT) were performed at the same hospital in every patient.

We directly measured the intrarectal temperature using a 4-point microthermocouple sensor (Yamamoto Vinita, LAS-4, Osaka, Japan). The sensor was inserted into the rectum at the level of the uterus during each HT treatment. Intrarectal temperature measurement was performed because vaginal or uteral insertion is more complex and less well tolerated by patients. In addition, a previous study that compared intra-tumour and intra-luminal temperatures during deep regional HT demonstrated that distributions of intrarectal temperature showed a similar pattern as intra-tumour and intra-vaginal temperatures in CC patients [30]. The temperature was automatically recorded each minute during HT. The thermal parameters of the cumulative equivalent min at 43 °C for the T90 (CEM43T90), T90, T25 and T50 were obtained on the basis of the intrarectal temperatures during all of the HT sessions. The Tx is an index temperature, which indicates when the temperature was reached or exceeded by x% of the intrarectal measurement points. The median Tx was defined as the median value of Tx during each HT session. The thermal dose parameter of CEM43T90 has been used extensively (and successfully) in clinical trials to assess the efficacy of heating [19,31–33]. It represents the thermal isoeffect dose expressed in cumulative equivalent min at a reference temperature of 43 °C based on the low end of the temperature distribution (T90). The CEM43T90 was calculated from the time-temperature data using the following formula:

When the temperature is higher than 43 °C, R = 0.5. When the temperature is lower than 43 °C, R = 0.25. In this formula, t_i is the time interval of the ith sample (t_i = 1.0 min). The temperature exceeding the temperature at 90% of the intrarectal measurement points during the ith minute is designated as T90i. The CEM43T90 is then used to convert each T90i into an equivalent time at 43 °C. These equivalent times are added together over the entire treatment duration of n min.

The evaluation of the treatment outcomes

A complete response (CR) was achieved when no tumour was detected in a physical examination or by magnetic resonance imaging with negative cytology or a negative biopsy result for at least one month after treatment. OS was defined as all-cause death as an outcome event. DFS was all-cause death, local relapse or distant metastasis as outcome events. LRFS was local relapse as an outcome event. In the patients with non-CR of the tumour, the day of the local relapse was defined as the day of the completion of study treatments for the DFS and LRFS. A modification of the Radiation Therapy Oncology Group (RTOG) morbidity scale was used to score patient acute toxicity [34]. The patients were followed up at one month intervals during the first year and at two month intervals thereafter.

Statistical analysis

The LRFS, DFS, and OS rates were calculated using the Kaplan–Meier method. The statistical significance of the differences between the actuarial curves was assessed using the log-rank test. Univariate analyses were performed using certain factors to identify the prognostic factors for LRFS, DFS and OS. Multivariable Cox proportional hazards models were used to investigate the survival outcomes. The Mann–Whitney U-test was used to evaluate differences in the CEM43T90 values of the patients in the CR and non-CR groups. The CR rates and the frequency of acute-phase treatment-related toxicities between the groups of the patients with higher CEM43T90 (≥1 min) and lower CEM43T90 (<1 min) were analysed using Fisher’s exact test. A p < 0.05 was considered significant for all statistical tests.

Results

Figure 1 shows the thermal parameters in each of the HT treatment sessions for the 47 patients who received CRT + HT. Throughout the entire treatment course, the CEM43T90 ranged from 0.1 to 46.6 min (median, 1.1 min), the median T25 ranged from 39.0 to 43.0 °C (median, 40.6 °C), the median T90 ranged from 37.7 to 42.2 °C (median, 38.9 °C) and the median T50 ranged from 38.4 to 42.4 °C (39.9 °C).

Figure 1. The thermal parameters (CEM43T90, median T25, median T90 and median T50) in each of the HT treatment sessions.

The five-year LRFS rate of the 47 patients was 81%. Among these patients, a univariate analysis showed that higher CEM43T90 (≥1 min) values over the whole course of treatment and histology of squamous cell carcinoma (SqCC) were significantly associated with better LRFS (p = 0.024 and p = 0.034) (Table 3 and Figure 2). In subsequent multivariate analyses, higher CEM43T90 (≥1 min) and SqCC were both significantly associated with a better prognosis (p = 0.0097 and p = 0.0046) (Table 3). LRFS rates for the 26 patients with a higher CEM43T90 (≥1) and the 50 patients treated with CRT alone were significantly different (p = 0.036), whereas LRFS rates for the 21 patients with a lower CEM43T90 (<1) and the 50 patients treated with CRT alone were not different (Figure 2).

Figure 2. (a) The LRFS for the 26 cases with CEM43T90 ≥ 1 min was significantly better than LRFS for the 21 cases with CEM43T90 < 1 min (p = 0.024) and the 50 CRT alone cases (p = 0.036). The CEM43T90 < 1 min cases had LRFS indistinguishable from the CRT alone cases (p = 0.80). At five years the LRFS for the CEM43T90 ≥ 1 min, CEM43T90 < 1 min, and CRT alone cases were 92%, 65%, and 71% respectively. (b) The DFS for the cases with CEM43T90 ≥ 1 min was significantly better than DFS for the CRT alone cases (p = 0.036), but did not achieve statistical significance when compared with the CEM43T90 < 1 min cases (p = 0.071). The CEM43T90 < 1 min cases had DFS indistinguishable from the CRT alone cases (p = 0.89). At five years the DFS for the CEM43T90 ≥ 1 min, CEM43T90 < 1 min, and CRT alone cases were 81%, 60%, and 61% respectively.

Table 3. The factors predicting survival in the 47 patients treated with definitive CRT plus regional HT.

		LRFS				DFS		OS
		Univariate		Multivariate		Univariate		Univariate
Variable	n	3-y (%)	p	OR (95% CI)	p	3-y (%)	p	3-y (%)	p
Age
≥65	20	90	0.18	–	–	85	0.43	89	0.5
<65	27	73				69		77
Stage
IIb	11	91	0.35	–	–	82	0.54	91	0.6
IIIB, IVA	36	77				74		79
Tumour size (cm)
≥5.5	23	74	0.24	–	–	70	0.17	76	0.4
<5.5	24	87				83		87
Histology
SqCC	40	85	0.034	11.8 (2.1–66.6)	0.0046	79	0.13	86	0.1
Non-SqCC	7	57				57		57
Total dose of cisplatin (mg/m^2)
200	30	76	0.34	–	–	69	0.13	75	0.2
<200	17	88				88		94
CEM43T90
≥1	26	92	0.024	11.5 (1.8–71.4)	0.0097	89	0.071	87	0.4
<1	21	65				60		75

LRFS: local relapse-free survival; DFS: disease-free survival; OS: overall survival; OR: odds ratio; SqCC: squamous cell carcinoma.

The five-year DFS rate of the 47 patients was 73%. A univariate analysis revealed that higher CEM43T90 (≥1 min) values tended to predict better DFS (p = 0.071) (Table 3). DFS rates for the 26 patients with a higher CEM43 °CT90 (≥1) and the 50 patients treated with CRT alone were significantly different (p = 0.036), whereas DFS rates for the 21 patients with a lower CEM43°CT90 (<1) and the 50 patients treated with CRT alone were not different (Figure 2).

The OS rate in the 47 patients was 65%. The OS rates of the 26 patients with higher CEM43T90 (≥1 min) values and the 21 patients with lower CEM43T90 (<1 min) values did not differ to a statistically significant extent. There was also no significant difference between for the OS rates between the 26 patients with a higher CEM43T90 (≥1) and the 50 patients treated with CRT alone.

After the completion of the protocol treatment, a CR was achieved in 41 of the 47 patients (87%). A CR tended to be associated with the CEM43T90; the median CEM43T90 value was 1.3 min in 41 patients with a CR; in contrast, the median value in the 6 patients with a non-CR was 0.7 min (p = 0.056) (Figure 3). A CR was recognised in 25 (96%) of the 26 patients with higher CEM43T90 (≥1 min) values and 16 (76%) of the 21 patients with lower CEM43T90 (<1 min) (p = 0.076). CR rates for the patients with a higher CEM43°CT90 (≥1) and the patients treated with CRT alone were significantly different (p = 0.048), whereas CR rates for the patients with a lower CEM43°CT90 (<1) and the patients treated with CRT alone were not different.

Figure 3. The CR rate tended to be correlated with the CEM43T90 value (p = 0.056). The error bars show the standard error.

The incidence of acute-phase toxicities and the RTOG grades of the CEM43T90 (≥1 min) and CEM43T90 (<1 min) were similar (Table 4). There were no significant differences between the two groups in the incidence of any grade 3–4 acute-phase toxicities. The incidence of late-phase toxicities was also similar between the two groups. Late toxicities in the patients treated with CRT + HT with CEM43T90 (≥1 min) were as follows: ileus in 2 (8%) patients, vesico-vaginal fistula in one (4%) patient and leg oedema in one (4%) patient. The findings in the patients treated with CRT + HT with CEM43T90 (<1 min) were as follows: ileus in 2 (10%) patients, ureter stenosis in one (5%) patient and leg oedema in one (5%) patient.

Table 4. Acute-phase toxicities between the CRT + HT with CEM43T90 (≥1) and CRT + HT with CEM43T90 (<1).

	CRT + HT (CEM43T90: ≥1) (n = 26) (%)				CRT + HT (CEM43T90: <1) (n = 21) (%)
Variable	Grade 1	Grade 2	Grade 3	Grade 4	Grade 1	Grade 2	Grade 3	Grade 4
HGB	2 (8)	16 (61)	5 (19)	1 (4)	0	12 (57)	6 (29)	2 (10)
WBC	1 (4)	8 (31)	12 (46)	3 (12)	1 (5)	8 (38)	5 (24)	5 (24)
PLAT	6 (23)	4 (15)	2 (8)	0	6 (29)	1 (5)	4 (19)	0
Nausea	7 (27)	0	0	0	6 (29)	0	0	0
Vomiting	2 (8)	0	0	0	2 (10)	0	0	0
Diarrhoea	11 (42)	1 (4)	0	0	4 (19)	0	0	0
Urinary	4 (15)	0	0	0	3 (14)	1 (5)	0	0
Weight loss	3 (12)	0	0	0	1 (5)	1 (5)	0	0
Fatigue	5 (19)	0	0	0	8 (38)	1 (5)	0	0
Nephrotoxicity	5 (19)	1 (4)	0	0	5 (24)	6 (29)	0	0
Blistering	0	0	0	0	0	0	0	0

Discussion

The present study demonstrated that the thermal dose parameter was associated with better clinical outcomes in patients with CC of the uterus treated with CRT + HT. The strengths of the current study include the fact that it used prospectively randomised sampling data, and that its detailed treatment protocol was defined. Many clinical phase III trials have indicated that the addition of HT to RT can improve clinical outcomes, especially local control rates and CR rates, in patients with superficial tumours [20,35,36]. However, as noted previously, meta-analyses of the thermal data from four clinical phase III trials evaluating RT, with or without HT in the treatment of advanced and recurrent breast cancer, showed that significant improvements in local control rates were only achieved in the patients in whom higher intra-tumour temperatures were achieved [20]. In experimental studies using canine and feline tumours, the efficacy of CEM43 to predict tumour response or local control was confirmed during combined treatment with RT + HT [37,38]. Franckena et al. [21] also demonstrated a dose-effect relationship of HT in CC patients who were treated with definitive RT alone + deep regional HT; intraluminally measured thermal parameters, including CEM43T90, were significantly and independently correlated with tumour control and survival. Sreenivasa et al. [26] reported the positive relationships between certain thermal parameters and the tumour response in 32 patients with non-resectable CC who were treated with preoperative CRT + deep regional HT. They indicated that the CEM 43 °C values in the vagina of 24 responders were significantly higher in comparison to 6 non-responders. In the current study of definitive CRT + deep regional HT in patients with CC, a higher CEM43T90 value in the rectum was found to be significantly associated with better LRFS, and it also tended to be significantly associated with better DFS and CR rates. This study also confirmed that regional HT with lower CEM43T90 is insufficient to achieve a significant hyperthermic sensitisation to CRT.

Excessive power deposition in the subcutaneous fatty tissue, which is associated with pain or thermal burn, may limit the effectiveness of capacitive regional HT and there is also a depth limit in the skin-cooling ability of the 8 MHz RF capacitive heating device [39]. Yahara et al. [24] reported the thermal analyses of 75 prostate cancer patients who received RT + deep regional HT of the pelvis using the 8 MHz radiofrequency-capacitive heating device and indicated that higher CEM43T90 values in the rectum could be achieved in patients with thinner subcutaneous fat in the pelvis. In other studies of capacitive devices in which thermal analyses were performed, thinner subcutaneous fat, a good performance status, and a younger age were significant predictive factors for successful deep regional heating in the thoracic region [23,40]. Asian patients are considered to be relatively suitable for capacitive regional HT due to their slender constitution, whereas radiative phased array devices are more suitable for administering regional HT to Caucasian patients. The current study confirmed, based on the data from a prospective randomised trial, that higher thermal parameters significantly improved the LRFS of CC patients who were treated with definitive CRT + HT. Thus, the further development of deep regional HT methods should be promoted. At the same time, Jones et al. conducted a randomised, phase III trial in patients with superficial cancer treated with RT, with or without HT; patients received a test dose of HT and tumours deemed heatable (≥0.5 CEM43°CT90) were randomly assigned to receive additional HT versus no additional HT [32]. Future randomised trials evaluating the addition of regional HT to CRT in patients with CC should be designed to select patients who are more likely to benefit from HT.

Several prognostic factors have been reported in CC patients treated with CRT. A large population-based study recently showed that a histological finding of adenocarcinoma was still an independently associated with a worse prognosis in CC patients treated with concurrent CRT [41]. On the other hand, Rose et al. [42] reported, using the data from the Gynaecologic Oncology Group trials, that non-squamous cell carcinoma was associated with similar progression-free and OS to squamous cell carcinomas of the cervix when patients were treated with CRT using cisplatin. Parker et al. [43] performed a retrospective analysis and demonstrated that the tumour size was a significant predictor of OS after CRT in CC patients. Nugent et al. reviewed 118 patients with locally advanced CC who were treated with definitive RT and weekly cisplatin (40 mg/m²). They found that nearly 70% of the patients successfully completed six scheduled cycles of cisplatin, and that the completion of 5 or 6 cycles of cisplatin concurrently with RT was significantly correlated with PFS and OS [44]. In the current study, neither the tumour size nor the total dose of cisplatin was significant predictors of LRFS; however, the CEM43T90 value and the tumour histology were significant predictors of LRFS in both the univariate and the multivariate analyses. Thus, the CEM43T90 value might be an independent factor that predicted LRFS in patients treated with definitive CRT plus regional HT.

The present study was associated with several limitations. First, the complete thermal data were not available for four of the patients in the CRT + HT group in the JASHO randomised trial, and these patients were excluded from the analysis. Although the remaining 47 patients with complete thermal data could be properly analysed, the missing data might have influenced the results of the current study. Second, the small study population of the present study made it difficult to detect significance. The application of regional HT as a treatment for malignancies is currently limited throughout the world. Our findings should be validated in future phase III studies or meta-analyses involving larger study populations.

In summary, this is the first report to assess the dose-effect relationship between thermal dose parameters and efficacy of definitive CRT + regional HT in patients with locally advanced CC. The previously reported results of the prospective, multicentre, randomised, parallel-group study conducted at five Japanese institutions by the JASHO group indicated the only significant benefit of adding regional HT to CRT was an improvement of the CR rate [25]. In the current study, we demonstrated that the addition of regional HT to CRT could improve the clinical outcomes in patients with higher CEM43T90 values. Our results provide justification for further evaluations of this treatment strategy. The results also indicate the importance of the accurate selection of treatable patients with higher thermal parameters and that further benefits may be achieved by progress in deep regional heating methods.

Disclosure statement

No potential conflict of interest was reported by the authors.