Radiation-induced DNA double-strand breaks in peripheral leukocytes and therapeutic response of heel spur patients treated by orthovoltage X-rays or a linear accelerator

The successful treatment of benign inflammatory and degenerative conditions like calcaneodynia (painful heel spur) and painful shoulder or elbow syndrome with low-dose radiation therapy (RT) has a longstanding history in Germany and accounts for more than a third of all RT patients treated annually [1]. Local administration of photon doses below 1 Gy has been shown to attenuate inflammatory responses as the causative factor for these painful diseases [2]. Clinical studies documented comparable positive results for treatment schedules with single fractions of 0.5 Gy or 1 Gy applied twice a week for 3 or 6 weeks [3‐5]. Accordingly, the iso-effective single dose of 0.5 Gy over a course of 3 weeks is used for radiation protection purposes to reduce the radiation burden to the patient and the risk of potential radiation-induced (RI) stochastic late effects such as radiation carcinogenesis [6]. The risk of adverse late effects induced by low-dose RT for benign diseases is generally considered to be negligible and the benefit for the patient outweighs [7‐9]. Anthropomorphic phantom-based studies by Jansen et al. [10] using an effective dose concept for a cumulative dose of 6 Gy administered in six fractions of 1 Gy predict rough estimations for the risk of RI fatal tumors of 20–40 per 1000 patients for heterotopic ossification, 1.5 per 1000 patients for gonarthrosis, 0.5 per 1000 patients for heel spurs, and 1 per 1000 female patients for hidradenitis suppurativa. Other predominant factors that influence the risk of radiation carcinogenesis besides the target dose are the field size, photon energies, the exposed anatomic region/organs, sex, and age at exposure. In general, such risk estimates are fraught with large uncertainties in the order of a factor of 2 and are expected to be much less for low-dose RT of benign diseases, since this treatment is predominantly performed in elderly patients with a strongly and up to nine-fold reduced excess lifetime risk for RI solid tumors when compared to young adults [10‐12].

Calcaneodynia is observed in 10–15% of the population and is associated with severe heel pain causing restrictions and reduced quality in everyday life [13, 14]. In Germany, approximately 10,000 heel spur patients per annum are treated by low-dose RT in about 340 active facilities offering RT for benign diseases [15]. Treatments are usually conducted with a medical linear accelerator (linac) and megavoltage (MV) photons or orthovoltage devices delivering X‑rays in the range of 100–400 keV, with a prevalence for linacs [15]. Compared to an orthovoltage device, RT with a linac achieves a more homogeneous dose distribution in the target volume, which has been discussed to have improved therapeutic effectiveness [16]. However, the penetration depth of MV photons from a linac exceeds the cross-section of the patient and causes reflections in the RT bunker that contribute to an inevitable dose burden to the patient’s normal tissue outside the target volume. So-called out-of-field doses are also generated by scattering and leakage from the linac treatment head, collimation devices, and, to a lesser extent, by scattering from within the patient’s body [17]. Such low-dose exposures far from the primary beam have been associated with various radiation-related late adverse effects [18, 19]. Orthovoltage RT with low-energy photons has an inferior dose distribution but allows for protection of the patient’s healthy tissue by lead shielding which cannot be applied for high-energy photon units due to the generation of scatter radiation within the shielding equipment. However, for linac RT, treatment planning including field collimation can be conducted to protect the patients’ healthy tissue outside the target volume.

Besides conventional physical measurements to determine the patient’s radiation exposure, the detection of biological indicators of RI deoxyribonucleic acid (DNA) damage in peripheral leukocytes has been frequently applied for biodosimetric purposes in patients undergoing various radiologic procedures or RT [20‐29]. By far the most sensitive and rapid measure is the immediate quantification of RI DNA double-strand breaks (DSB) using the phosphorylated histone variant H2AX (γH2AX) or tumor protein 53 binding protein 1 (53BP1), representing well-established surrogate markers of DSBs [30]. Immunostaining and fluorescence microscopic quantification of so-called γH2AX or 53BP1 foci at the level of single cells is proportional to the number of RI DSBs and therefore increases linearly with radiation dose after ex vivo and in vivo exposure with a detection threshold of a few mGy [22, 26, 27].

However, it is hitherto unknown whether the largely random application of linac or orthovoltage RT for the treatment of painful heel spur is associated with a varying unwanted but inevitable radiation exposure of the patients’ normal healthy tissue to out-of-field doses and differing therapeutic effectiveness. Therefore, we conducted this prospective randomized trial as a biodosimetric approach to assess the radiation burden of heel spur patients treated either with orthovoltage X‑rays or a linac by quantitation of the DSB marker γH2AX in peripheral leukocytes. RT was given twice per week with a single dose of 0.5 Gy and six total fractions. Venous blood samples were drawn immediately before and 30 min after the first fraction of RT and were processed for fluorescence microscopic scoring of γH2AX foci in leukocytes. As a secondary endpoint the analgesic effectiveness was monitored based on a five-level function calcaneodynia score (CS) before and 3 months after RT.

In the present study we measured RI DSBs by γH2AX foci quantification in peripheral leukocytes of painful heel spur patients treated with a 140 kV orthovoltage device or a 6-MV linac to assess the patients’ radiation burden for radiation protection purposes. Immediately after the first fraction of RT we detected an overall slight but significant increase of γH2AX foci with no difference between orthovoltage and linac RT. The application of either RT technique led to significant and comparable pain relief at 3 months follow-up. Based on this outcome as well as low participation and inclusion rates, the trial was terminated preliminarily after an interim analysis (22 patients randomized).

DSBs are potently induced by ionizing radiation and represent the most deleterious DNA lesion causing cell death, chromosomal rearrangements, and malignant transformation [33]. The by far most prominent biomarker of RI DSBs is the phosphorylated histone variant γH2AX, which has been applied in numerous studies to evaluate the in vivo radiation exposure of patients after low-dose radiologic examinations like computed tomography (CT) (e.g., [20, 22, 23, 34]) and mammography [21, 25], or after high-dose RT of tumor patients [24, 26‐29]. The present study is the first, at least to our knowledge, to apply this method with a biodosimetric intention in patients treated by low-dose RT for benign inflammatory and degenerative diseases with pain relief as the second clinical endpoint. RT of this medical condition is very well received and frequently used in German-speaking regions but is barely applied in other, particularly Anglo-American, countries [1]. Such geographical differences are due to fear of RI late adverse effects, which, however, are estimated to be very low or negligible for this type of local low-dose RT [7, 8]. Various studies on the RT of heel spurs showed equal effectiveness for single doses of 1 Gy or 0.5 Gy administered twice a week for 3 or 6 weeks [3‐5, 16]. Accordingly, the lower dose of 0.5 Gy represents the standard option to decrease the potential risk for radiation-related late adverse effects [6]. Depending on the institutional equipment, benign diseases are treated either with a linac and MV photons or with an orthovoltage device operating in the low-energy kV range. This instrumentation-specific difference might be associated with varying therapeutic effectiveness [16] but also with divergent exposures to undesirable out-of-field doses [17]. Also, from a health economic perspective, orthovoltage RT is associated with lower costs compared to linac RT. Based on this rationale, we investigated both endpoints in this prospective randomized trial after low-dose RT of calcaneodynia patients with a linac or orthovoltage unit for a treatment schedule of 0.5 Gy given twice a week over a course of 3 weeks. According to our scoring criteria adapted from Rowe et al. [31] and Heyd et al. [3], we observed high response rates and pain reduction in up to 89% of patients at 3 months follow-up in line with improvement rates of previous studies ranging between 65–100% [16]. About 50% of patients reported on an excellent and pain-free performance status after the first RT series. We did not observe any difference in the therapeutic response between the two RT techniques, but the small number of participants does not allow meaningful statistical comparisons. Previously, Muecke et al. [16] performed a retrospective study on the long-term treatment success of low-dose RT for painful heel spurs in 502 patients treated either with 6–10-MV photons twice per week or with 175 kV X‑rays three times per week at four different facilities in Germany. Patients received 10 fractions of 0.5 Gy or 5–6 fractions of 1 Gy for 6–10-MV photons or six fractions of 1 Gy for orthovoltage X‑rays. In their study, multivariate analysis revealed a significantly worse prognosis for orthovoltage RT than for MV photons, with no impact of radiation dose. This finding has been related to a more homogeneous and favorable dose distribution achieved with MV units. No other study has yet confirmed this observation. Although a better distribution of dose is achieved in the target volume for linac RT, it may increase the radiation burden of the patient through higher peripheral doses outside the primary beam caused by radiation scattering, leakage, and reflections [17]. Besides physical dosimetry, biodosimetric attempts have been made to compare the inherent radiation exposure of different RT techniques [27‐29] or CT protocols [35] based on the quantification of RI γH2AX foci in peripheral leukocytes. Thresholds for this highly sensitive assay to monitor the induction of RI DSBs in vitro and in vivo have been set at 1 mGy and 3 mGy, respectively [22, 36]. So far, only few comparable studies on foci quantification of DSB repair proteins in systemic lymphocytes after a planned medical IR exposure in vivo are available for the low EWBDs of the present work, which were able to demonstrate dose-dependent increments or even differences between radiation techniques. Kuefner et al. [35] reported on significantly reduced levels ofγH2AX foci in peripheral leukocytes 30 min after multidetector coronary CT angiography performed with a dose-reducing sequential protocol compared to a conventional helical protocol in line with physical dose estimates. For an approximated median effective dose ranging from 2.1 to 23.8 mSv, the authors described a linear dose response for the induction of excess γH2AX foci in vivo from 0.04 to 0.71 foci per leukocyte with a median of 0.33 in line with similar studies [22]. In another study these authors investigated the impact of digital mammography executed with doses even lower than for CT examinations [25]. Again, a very slight but significant increment of γH2AX foci was found in systemic leukocytes of 20 patients. The average EWBD in our study was estimated to be in the range of just 2.34–14.67 mSV and was significantly higher for linac than for orthovoltage RT. Based on our calculations, this variation of the EWBD between the radiation techniques was determined by differences in the SSD and half-value thickness. For otherwise identical parameterization, a higher SSD for linac RT caused an average 1.9-fold reduction of the integral dose compared to orthovoltage RT and, conversely, the higher half-value thickness for linac RT resulted in a six-fold increment of the integral dose than for orthovoltage RT. The impact of these two parameters resulted in a general significant 4.4-fold increase of the integral dose for linac RT compared to orthovoltage RT. This value also applied for calculated EWBD, since there were no significant differences in the distribution of the patients’ bodyweight between the two radiation modalities. We observed a general slight but significant increase of γH2AX foci per cell after RT for all patients but no difference between the RT techniques nor a correlation with the EWBD. 30 min after RT the numbers of excess γH2AX foci per cell were in a low range of 0 to 0.685, with a median of 0.104.

In our previous studies on the quantification of γH2AX foci in peripheral leukocytes of breast and prostate cancer patients after RT, we have shown linear dose–response relationships and good approximations of the administered whole-body dose based on ex vivo calibration data [26, 27]. However, cancer patients were exposed to significantly higher EWBDs compared to patients with benign diseases of the present study (Fig. 3d). According to our reference data on linear dose–response relationships of γH2AX foci in leukocytes at various times post exposure [26], the average frequency of 0.149 RI foci per cell after RT of all patients in the present work equates to a mean absorbed X‑ray dose of 15.1 mGy, which exceeds our calculated average EWBD of 6.77 mSv. But assuming that the number of foci of DSB repair proteins in peripheral leukocytes after RT is a quantitative measure of the patient’s dose burden and correlates with the risk of adverse side and late effects of medical radiation exposure, it is expectedly and significantly far lower in heel spur patients than for tumor patients. However, as we reported in our previous work [26, 27], the induction of DSBs in peripheral leukocytes during RT depends on various radiation and physiological parameters, which strongly limits such direct comparisons. Although the yield of γH2AX foci in peripheral leukocytes during RT of cancer patients is primarily governed by general RT parameters such as the planning target volume or the administered EWBD, we described volume- and dose-independent variations of radiation biomarkers in leukocytes among different RT techniques for breast cancer treatment which were heavily dominated by the absolute beam-on time [27]. Therefore, the radiation parameters between the two RT techniques of the present study, such as the field size or the dose rate, were adjusted as well as possible to achieve comparable exposure scenarios and beam-on times, to detect the impact of diverse out-of-field doses only. Besides, a strong dependency of foci induction in systemic leukocytes on physical variables such as the regional blood volume and kinetics of leukocyte circulation in the exposed anatomic region has to be considered for any comparative tactic [22, 24]. These confounding factors also greatly deteriorate the accuracy of radiation biomarkers for dose estimates after RT, in particular in the range of very low doses.

Taken together, using a biodosimetric approach to monitor the radiation burden of heel spur patients after the first fraction of RT with a single dose of 0.5 Gy administered with a 140-kV orthovoltage device or a 6-MV linac, we observed a marginal but significant overall increase in the DSB surrogate marker γH2AX in peripheral leukocytes, with no difference between the RT techniques. Both treatment modalities were associated with very modest radiation exposures and showed high and comparable analgesic effectiveness. Our data confirm the use of low-dose RT as an attractive treatment option for benign diseases.