Local Control and Toxicity of External Beam Reirradiation with Pulsed Low-Dose-Rate Technique

Abstract

Purpose:

To evaluate the efficacy and toxicity of external beam reirradiation using pulsed low-dose-rate (PLDR) technique.

Methods/Materials:

We evaluated patients treated with PLDR reirradiation between 2009–2016 at a single institution. Toxicity was graded based on CTCAE4.0 and local control was assessed using RECIST1.1. In the univariate analysis (UVA), Chi-square, and Fisher’s Exact test were used to assess toxicity outcomes; competing risk analysis via cumulative incidence function estimates were used to assess local progression.

Results:

Thirty-nine patients were treated to 41 disease sites with PLDR reirradiation, with a median follow-up of 8.8 months (range 0.5–64.7 months). Targets were thoracic, abdominal and pelvic, including 36 symptomatic sites. The median time between the first radiation course and reirradiation was 26.2 months; the median doses of the first and second courses of radiation were 50.4 Gy and 50 Gy, respectively. Five patients (13%) received concurrent systemic therapy.

Out of 39 patients, 9 (23%) developed grade 2+ acute toxicity, most commonly radiation dermatitis (5/9). None developed grade 4+ acute/subacute toxicity. The only grade 2+ late toxicity was 1 patient with grade 2 late skin toxicity. In UVA, toxicity was not significantly associated with dose of the first course of radiation or reirradiation, time interval to reirradiation, or reirradiation site.

Of the 41 disease sites treated with PLDR, 32 had pre and post-PLDR scans to evaluate for local control. Local progression was 16.5% at 6 months and 23.8% at 12 months, and was not associated with the dose of reirradiation, reirradiation site, or concurrent systemic therapy in UVA. Of 36 symptomatic disease sites, 25 (69%) sites achieved symptomatic response after PLDR, including 6 (17%) with complete symptomatic relief.

Conclusion:

Reirradiation with PLDR is effective and well-tolerated. Risk of late toxicity and durability of local control was limited by the relatively short follow up in this cohort.

SUMMARY

Severe acute and late toxicities are a barrier to conventional reirradiation in patients with recurrent tumors or tumors arising in sites previously treated. Our study retrospectively evaluates the efficacy and toxicity of reirradiation with pulsed low-dose-rate technique which is a novel way to deliver reirradiation. Results from our study suggest that it can be safe technique to deliver reirradiation and limit the development of severe toxicities, although our results are limited by relatively short patient follow-up.

INTRODUCTION

Demographic and disease-related factors such as increasing life expectancy and earlier cancer detection have led to an increasing number of cancer patients to require treatment with radiation [1, 2]. In addition, patients with metastatic cancer are living longer due to advances in systemic therapy [3]. The need for increased durability of treatment in the context of a persistent risk of tumor recurrence in a treated region has resulted in an increasing need for radiation to volumes that were previously irradiated. Reirradiation, however, has been limited historically due to concern for serious late toxicities. The cumulative dose necessary to achieve a significant tumor response may exceed the tolerance of surrounding normal tissues, while the upper-limit of tolerable dose in the reirradiation setting is largely unknown [4, 5]. Thus, reirradiation is typically a last resort in patients who are not candidates for other therapies. For some cancer patients, the risk/benefit ratio of reirradiation becomes more favorable when a tumor in a previously irradiated field causes significant symptoms, such as bronchial obstruction, spinal cord compression, or intractable pain, that cannot be alleviated by other modalities [6]. For such patients receiving palliative treatment, the ability to deliver reirradiation safely and effectively is imperative because the alleviation of symptoms must be more substantial than the side effects of reirradiation for the treatment to be worthwhile.

Pulsed low-dose-rate (PLDR) external beam radiation is a technique of delivering radiation by breaking conventional radiation doses into pulses of smaller doses with discrete time intervals in between [7–9]. When the dose rate is <100 cGy/minute, tumor cells exhibit hyper-radiosensitivity while normal tissues undergo improved sublethal damage repair [10–14]. As such, the therapeutic index of reirradiation is enhanced which allows improved tumor kill relative to side effects. By breaking up a conventional fraction of radiation into several small pulses, PLDR is a clinically practical way to achieve low dose rates in a reasonable time period.

Few human studies with PDLR have been published to date. Of the limited evidence available, PLDR has been shown to be potentially effective for palliative retreatment of gliomas and breast cancers [7, 8, 15]. Data for other types of cancer are sparse. At our institution, irradiation with PLDR has been delivered using conventional techniques as well as intensity-modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT) since 2006, harnessing the technological advances used in conventional treatment to further reduce the risk to nearby critical structures [9, 16–18]. In this retrospective study, we report the toxicity and efficacy of reirradiation with PLDR.

MATERIALS AND METHODS

We reviewed patients who received reirradiation with PLDR technique at our institution from 2006 to 2016. PLDR was typically delivered 5 days per week with daily fractions of 2 Gy in 10 pulses of 0.2 Gy with 3 minute separation from the beginning of two adjacent pulses, using IMRT, VMAT, or 3D conformal therapy [9, 16–18]. In general, doses delivered to critical structures were kept as low as possible. However, conventional constraints were frequently exceeded by the cumulative doses of first and second courses of radiation.

Toxicity was graded using Common Terminology Criteria for Adverse Events (CTCAE) v4 [19]. We defined toxicity that developed during radiation or within 90 days post radiation as acute, and after 90 days post radiation as late toxicity. Radiation toxicity evaluated in this study included radiation dermatitis, fatigue, anorexia, nausea, dysphagia or odynophagia, cough, and other rare events that could be potentially attributed to radiation toxicity. Among patients with imaging studies taken before and after PLDR, disease response was evaluated using Response Assessment in Solid Tumors (RECIST) v1.1 criteria if there was measurable disease [20–22]. Local failure was defined as progressive disease per RECIST v1.1 criteria. Patients that did not have imaging to assess local control following PLDR were not included in this assessment.

For univariable analysis (UVA), Fisher’s Exact model was used to test the associated factors for their effects on acute and late toxicity: concurrent systemic chemotherapy, irradiation of bone metastases, location of reirradiation, patient age, time interval between first RT and PLDR, dose of first RT, and dose of PLDR. The Cumulative incidence function (CIF) estimation method was used to test the associated factors for local recurrence and local control after controlling for the competing risk of death. Overall survival was also assessed via the Kaplan-Meier method. Multivariable analysis was not done due to the small number of events.

RESULTS

There were 39 patients (18 male and 21 female) treated to 41 disease sites with PLDR reirradiation from 2006–2016 at our institution. The median follow-up was 8.8 months (range 0.5–64.7 months) and the median time between the first course of radiation and reirradiation was 26.2 months (range 5.4–115.0 months). The median dose of the first course of radiation was 50.4 Gy (range: 24 Gy in 3 fractions to 140 Gy via low dose rate BT), and the median dose of PLDR reirradiation was 50 Gy (range: 28 Gy in 14 fractions to 60 Gy in 30 fractions). A total of 34 (87%) patients received systemic chemotherapy prior to reirradiation, and 5 (13%) continued chemotherapy during their PLDR treatment; 5 (13%) never received chemotherapy prior to or after PLDR. The location of PLDR reirradiation was thorax in 25 (61%), abdomen in 3 (7%), and pelvis in 13 (32%). There were 8 (21%) patients treated for bone metastases. The vast majority of patients (87%) received PLDR reirradiation for palliation. Thirty-six sites treated with PLDR reirradiation were symptomatic, including 21 tumor sites causing pain. The detailed patient characteristics are listed in Table 1.

Table 1:

Patient characteristics

Total number of patients (N)		39
Number of disease sites reirradiated with PLDR (n)		41
Age: median (range)		62 (42–86)
Sex	Male: N (%)	18 (46.2%)
	Female: N (%)	21 (53.8%)
Time between two courses of RT (month): median (range)		27 (5–115)
First RT dose (Gy): median (range)		50 (24–80)
PLDR dose (Gy): median (range)		50 (28–60)
Intent of reirradiation	Palliative: N (%)	34 (87.2%)
	Curative: N (%)	5 (12.8%)
Reirradiation region	Thorax: n (%)	25 (61.0%)
	Abdomen: n (%)	3 (7.3%)
	Pelvis: n (%)	13 (31.7%)
Presenting symptom	Pain: n (%)	21 (51.2%)
	Cough/hemoptysis: n (%)	8 (19.5%)
	Swelling/edema: n (%)	3 (7.3%)
	Dysphagia: n (%)	1 (2.4%)
	Other: n (%)	2 (4.8%)
	None: n (%)	6 (14.6%)

Abbreviations: RT-radiotherapy

Note: N is percent calculated based on total patients, n is percent calculated based on total disease sites

Of the 41 disease sites in 39 patients treated with PLDR reirradiation, 32 disease sites were imaged before and after PLDR reirradiation with the same modality, and were therefore evaluable for RECIST 1.1 assessment of local control. Local progression was 16.5% at 6 months and 23.8% at 12 months. There were 32 disease sites initially at risk based on pre- and post-PLDR imaging; there were 17 and 7 disease sites at risk at 6 and 12 months, respectively (Figure 1). Local control was not significantly associated with the dose of reirradiation, reirradiation site (thoracic, abdomen, or pelvis), or concurrent systemic therapy in UVA. Of the 36 symptomatic disease sites, 25 (69%) sites achieved clinical response after PLDR, including 6 (17%) with complete symptomatic relief. For those 21 tumor sites causing a chief complaint of pain, pain was improved in 14 (67%). For overall survival, 4 living patients were lost to follow-up prior to 6 months; of the remaining 35 patients, 25 (71%) were alive at 6 months. An additional 2 living patients were lost to follow-up prior to 12 months; of the remaining 33 patients, 13 (39%) were alive at 12 months (Figure 2).

Figure 1:

Local control for disease sites reirradiated with PLDR

Figure 2:

Overall survival of patients reirradiated with PLDR

Out of 39 patients treated with PLDR reirradiation, 30 (77%) developed any acute toxicity, including 11 (28%) with grade (G)2+ acute toxicity, most commonly radiation dermatitis (5/11). Grade 3 acute toxicity occurred in 5 (13%) patients of which 3 patients experienced radiation dermatitis, 1 patient experienced esophagitis, and 1 patient experienced lung fibrosis. None developed G4+ acute toxicity (Table 2). There were 3 patients with any late toxicity, and the only late G2+ toxicity was 1 patient with G2 late skin toxicity (Table 2). In UVA, acute and late toxicities were not significantly associated with dose of the first course of radiation or reirradiation with PLDR, time interval to reirradiation, patient age, reirradiation site, concurrent systemic chemotherapy, or presence of bone metastases.

Table 2:

Acute and late toxicities in patients treated with PLDR reirradiation

		G1: n	G2: n	G3: n
Acute toxicity	Thorax (n=25)	12	4	4
	Abdomen (n=3)	2	0	0
	Pelvis (n=13)	6	0	1
Late toxicity	Thorax (n=25)	1	0	0
	Abdomen (n=3)	0	0	0
	Pelvis (n=13)	1	1	0

Abbreviations: G1-grade 1; G2-grade 2; G3-grade 3

DISCUSSION

The potential to exceed normal tissue tolerance and cause significant patient harm has been a major limitation on the clinical use of reirradiation. Limited series investigating reirradiation using conventional technique demonstrates that while local disease control is often sufficient, conventional reirradiation can come with the cost of significant toxicities. Reirradiation to the head and neck region using conventional fractionation and technique has been associated with fatal bleeding and organ necrosis [23], severe brain necrosis, pharyngocutaneous fistulas, palatal fibrosis, osteoradionecrosis [24], and catastrophic carotid blowout [25]. Studies of thoracic retreatment resulted in severe esophagitis, G3 radiation pneumonitis in 5–21% of patients, G3 late pulmonary fibrosis in 9% patients, and fatal complications in 4% of patients [26–33]. Combined chemoradiation for recurrent rectal cancer led to a 15% incidence of small bowel obstruction requiring hospitalization and G3 chronic severe diarrhea in 17% of patients [34]. Mohiuddin, et al., utilized accelerated hyperfractionation for recurrent rectal cancer patients with a 34% rate of late G3–4 toxicity and a non-significant trend toward increased risk in patients who received a higher dose in the first course of radiation and a shorter interval to reirradiation. Subsequent surgery following reirradiation was associated with a significantly higher rate of complications such as wound infections, non-healing, and abscess formation [35]. Successfully retreated patients with gynecologic cancers experienced a 50% rate of severe complications such as chronic radiation cystitis, rectovaginal fistulas, and small bowel obstruction requiring surgery [36]. In the upper abdomen, one trial reported a 15% rate of treatment-discontinuing acute toxicities such as GI bleeding from an anastomosis and stricture formation [37].

In the present study, reirradiation using PLDR was well-tolerated with no interruptions for toxicity. Among 39 patients included in this study, only 5 (12.8%) developed G3 acute toxicity, the majority of them radiation dermatitis; none had G4 or higher acute toxicity. Only one patient had G2 late toxicity, and there were no G3 or higher late toxicities. No patients in our study required hospitalization due to PLDR toxicity. This late toxicity rate compares favorably to what has been reported in the studies described above. Additionally, our study demonstrates efficacy for local control at 6 and 12 months and symptom improvement. In patients with disease not amenable to surgery, PLDR was able to alleviate symptoms of pain, coughing, and dysphagia in 69% (25/36) of sites treated, with 17% (6/36) reporting complete symptomatic relief.

While PLDR has not been directly compared to standard reirradiation techniques, other published studies have supported its effectiveness in symptom and tumor control with favorable toxicity outcomes. Adkison et al.’s study of recurrent gliomas treated with PLDR resulted in no toxicities warranting treatment discontinuation [7]. Four of 15 patients (27%) who underwent brain biopsy showed notable necrotic areas but none were symptomatic. Additionally, 2-year survival from initial diagnosis for patients with grade 4 gliomas in the Adkison study was favorable at 21.7%; median survival from initiation of PLDR was 5.1 months. In a case report on metastatic lung cancer to the brain, complete resolution of headaches, vomiting, and dizziness was observed for a patient who underwent whole brain radiation and multiple doses of stereotactic radiosurgery prior to PLDR [38]. Additionally, Richards et al. report 15 of 17 (88%) patients receiving reirradiation with PLDR had complete resolution of their breast cancer recurrence at the primary site. Local control at 2 years was 92%. Toxicity profile was also acceptable; 4 patients experienced G3 acute toxicity and only 2 patients experienced late G3–4 toxicity in the form of non-healing chest wall ulcers [8].

Brachytherapy-based low-dose-rate approaches have also demonstrated favorable outcomes. Puthawala et al. demonstrated a 51% 5-year local control rate, overall, for patients with recurrent head and neck cancer. Although 56% of patients experienced soft-tissue necrosis and osteonecrosis, spontaneous healing or resolution following hyperbaric oxygen therapy occurred in 81% of these patients [39]. A similar approach by Strnad et al. reported an impressive 5-year local control rate of 82%; 28% of patients overall developed late toxicities and less than a quarter of these required surgical treatment [40]. Favorable outcomes were also seen for re-irradiated esophageal cancer patients in a study by Harms et al. A median 17-month symptom-free survival was accompanied by a 6% late toxicity rate after pulsed dose rate reirradiation with brachytherapy [41].

Our study is limited as a retrospective analysis without a head to head comparison with conventional retreatment. Toxicity may also be underreported owing to the retrospective nature of our study and the relatively short follow-up. Most patients had terminal disease, possibly contributing to underestimation of late toxicity due to short follow up. In addition, a diverse range of tumors and treatment sites are included in our study, which possibly affects both local control and toxicity outcomes compared to other literature which focus on specific anatomic locations or systems.

Nevertheless, the favorable late toxicity profile we found utilizing PLDR in the reirradiation setting is encouraging. A strength of our study is the relatively large number of patients treated with PLDR, and the assessment of tumor response by RECIST criteria. These data can provide an initial foundation of knowledge to safely offer reirradiation. While the ability to draw conclusions regarding the effectiveness of PLDR by specific histologic type is limited by the heterogeneity of the patient population, our trial has the advantage of demonstrating the generalizability of PLDR reirradiation as a safe and potentially effective option for patients who have no further systemic or local therapies available to them.

CONCLUSION

Reirradiation using PLDR technique appears to be a safe and effective treatment in palliation of symptoms from tumors in the thorax, abdomen, and pelvis. However, evaluation of late toxicity was limited by the relatively short follow up in this cohort of patients with mostly terminal disease. Prospective study is warranted.

Table 3:

Dose to regions experiencing G2+ acute and late toxicities

Patient Age	Gender	Tumor origin	Location of recurrence	Toxicity experienced	PLDR dose (Gy)	Maximum point dose to affected region (Gy)	CTCAE toxicity grade
49	F	Breast	Chest wall	Acute dermatitis	36	60	3
67	F	Lung	Supraclavicular nodes	Acute dermatitis	40	35	2
74	F	Breast	Chest wall	Acute dermatitis	45	50	2
50	F	Chest wall	Chest wall	Acute dermatitis	60	51	3
68	F	Lung	Mediastinal nodes	Acute esophagitis	30	28.17*	2
58	M	Rectum	Pelvis	Acute dermatitis	28	21	3
62	F	Rectal	Lung and mediastinal nodes	Acute pneumonitis	50	10.82* (V20=19.2%)	3
73	F	Lung	Subcarinal nodes	Acute esophagitis	60	10.94* (V55=8.18%)	3
69	F	Lung	Lung and mediastinal nodes	Acute chest wall inflammation	30	30	2
46	M	Rectum	Pelvis	Late dermatitis	60	20	2

Abbreviations:

^*Mean dose to entire affected organ

V20: Percent of affected organ receiving ≥20 Gy

V55: Percent of affected organ receiving ≥ 55 Gy