INTRODUCTION

To improve clinical outcomes for COVID-19 patients, a variety of experimental treatments are undergoing clinical trial evaluation, including whole-lung low-dose radiation therapy (LDRT) (1, 2), defined here as 0.3–1.5 Gy delivered in a single fraction, which is lower than doses used in clinical radiotherapy and higher than the 50 mSv/year occupational exposure limit and 1 mSv/year limit to the general public ( https://bit.ly/3ctWAAz).

Clinical trials currently testing whole-lung LDRT for COVID-19 are listed in Table 1. Most of these trials are open-label, single-institution, non-randomized studies, but some randomized trials are also emerging. Among several hospital centers worldwide, enrollment of 982 patients is planned to date. To address the uniqueness and accompanying scientific controversy of this unconventional treatment (3, 4), the National Cancer Institute (NCI), National Council on Radiation Protection and Measurements (NCRP) and National Institute of Allergy and Infectious Diseases (NIAID) convened a virtual workshop on July 23, 2020 to discuss available data concerning potential risks/benefits of LDRT. One main goal of the workshop was to establish a set of suggested guidelines for future clinical trials by developing components for inclusion in clinical studies, taking into consideration pros and cons of several parameters that would address the pathogenesis and underlying mechanisms of COVID-19.

TABLE 1

Completed and Ongoing Clinical Trials on Whole-Lung LDRT for COVID-19 Registered at ClinicalTrials.gov

The workshop brought together a diverse group of scientific experts in radiation oncology, pulmonary virology, immunology, vascular- and radiation-biology, radiation protection, and public health policy. There were three sessions: 1. Clinical trials/trial design; 2. Preclinical studies; and 3. Radiobiological/immunological mechanisms of LDRT. The available results of clinical trials conducted to date, including those just accruing patients and those being implemented and in the final planning phase, served as a key focal point of debate about the need for preclinical data prior to initiating trials and the protocol design and methodology. Issues discussed included those related to COVID-19 biology, low-dose radiation biology, radiation lung injury and how these complex mechanisms might interact. Given the use of ionizing radiation for a purpose other than treating cancer, potential life-time radiation risks and regulatory aspects of the use of radiation for LDRT were discussed. Without supporting or refuting LDRT for COVID-19 as a treatment, this review is intended to provide guidance to clinicians and researchers who plan to conduct preclinical and clinical studies, given the limited evidence on its safety and efficacy. Also, we recognize the need for novel treatments for patients in a dire clinical condition; however, it is essential that any use of LDRT for COVID-19 be on a clinical trial with Institutional Review Board (IRB) oversight, and investigators must provide complete details of trials, including patient selection, tracking those who were eligible, consented and received LDRT, and report both short- and long-term results.

CLINICAL TRIALS AND TRIAL DESIGN

The workshop participants recognized the dire need for effective treatments as well as scientific consensus from the NCRP ( https://bit.ly/3mS0AQ5), the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR;  https://bit.ly/3hZVg9L) and others, that radiation doses > 0.1 Gy carry long-term risks, which include cancer and cardiovascular events, modulated by several factors such as age at time of exposure and sex (3, 5, 6).

The impetus for LDRT comes from case studies of patients treated with thoracic irradiation for pneumonia between 1905 and 1943, which suggest clinical improvement (7, 8). As the studies lacked a randomized control arm without radiation exposure, it is difficult to establish whether treatment influenced the course of the disease. Given the paucity of preclinical data and limited interpret-ability of the historical data, some workshop participants insisted on the need to establish the efficacy of LDRT in animal models of SARS-CoV-2 before testing in human clinical trials, opining that LDRT is not indicated (9). Others argued for continuation of IRB-approved clinical trials based on the need to investigate treatments that aim to reduce COVID-19 morbidity and mortality and on initial phase I trial observations. While several other COVID-19 therapies are being evaluated, to date, only two drugs have shown improved outcomes: the steroid dexamethasone lowers the odds of death (10), and the antiviral remdesivir shortens oxygen dependency and recovery time but does not improve survival (11). Ultimately, management of this pandemic will be modified by improved understanding of pathogenesis of tissue injury, immunological/inflammatory response, the availability of preventive and therapeutic vaccines, and continually updated clinical guidance ( https://bit.ly/3kO4I1M).

The risk/benefit considerations for LDRT include the possibility of short-term worsening of the disease course (to be determined by phase 1 and 2 clinical trials) and long-term population-based radiation risks should a large number of patients live many years/decades vs. the potential benefit of reduced mortality and morbidity in the most severely ill patients in the immediate term (days/weeks). While speculative, it is conceivable that whole-lung LDRT could exacerbate SARS-CoV-2 infection. Autopsies of patients who died from COVID-19 show infection of endothelial cells with associated endothelial inflammation, vasculopathy and microangiopathy (12, 13). In general, a >2 Gy dose is likely to induce endothelial activation (14). Endothelial molecules, ICAM-1 (15) and E-selectin (16), are upregulated in a dose- and time-dependent manner, in part due to NF-κB activation (17), in preclinical studies. Because the endothelial cells in small vessel capillaries in the lung can potentially be damaged by ionizing radiation (14), it is conceivable that LDRT might promote endothelial cell injury and coagulopathy in the COVID lungs due to combined injury. Furthermore, there might be differences in the dynamics of response of endothelial cells between preclinical models and humans to a given dose of radiation. Therefore, characterization of dose-effect and time-course responses of endothelial cells to radiation are essential to fully understand the adverse effects of LDRT.

The general rationale for LDRT is that it could inhibit the cytokine storm that promotes the virus-induced pulmonary dysfunction contributing to the development of acute respiratory distress syndrome (ARDS). It is also likely that the potential anti-viral effect of LDRT could be due to the activation of immune and endothelial cells and inhibition of subsequent viral loading. However, preclinical data are necessary to support or refute this notion. LDRT as a treatment for COVID-19 patients to prevent respiratory failure, based on its role in treating inflammatory and degenerative diseases, has been reviewed elsewhere (18). Use of LDRT requires meticulous planning to overcome the challenges, including: 1. Determining the optimum time, if any, during disease progression to treat COVID-19 with LDRT, considering its safety and efficacy; 2. Safety of healthcare personnel during patient transport to the radiation therapy facility; 3. Rapid deterioration of eligible and enrolled patient's medical condition; and 4. Availability of radiation therapy facilities only on certain days (in many centers), potentially causing a delay between obtaining informed consent and the actual treatment delivery. Patients who deteriorate during this delay and fail to receive LDRT influence the interpretation of the existing preliminary phase 1 data.

The first published peer-reviewed results from the U.S. are from a safety trial of 1.5 Gy LDRT in elderly patients (median age 90 years), with bilateral pulmonary infiltrates, oxygen dependence without mechanical ventilation, and high comorbidity burdens, factors known to predispose patients to a worse outcome (2). A pre-specified interim safety analysis on day 7 was the primary end point: seven patients were enrolled (however, one was transferred to the ICU and one expired prior to LDRT); thus, five received radiation treatment. No acute worsening of the cytokine storm nor acute events were noted in four patients whose clinical status improved over 3 to 96 h post-LDRT. The trial met the primary safety end point with day 7 survival in the five LDRT-treated patients.

In Iran, five patients with COVID-19 on supplemental oxygen received 0.5 Gy whole-lung radiotherapy: one patient died, one withdrew consent and three showed clinical improvement so that they could be discharged, with one patient discharged with oxygen support at home (1). As these patients were not treated with dexamethasone, the death of one of four evaluable patients from COVID-19 is consistent with the expected mortality of 25% for hospitalized patients on supplemental oxygen (19). The authors reported a response rate of 4 out of 5 initially enrolled patients, based on an initial improvement in oxygen saturation and body temperature within one day after LDRT. Collectively, these studies established the feasibility of 0.5 to 1.5 Gy whole-lung irradiation in COVID-19 patients on supplemental oxygen.

Two trials underway that were discussed are: 1. A phase 2, single-arm study of 30-day mortality for patients requiring mechanical ventilation (NCT04427566) to receive 0.8 Gy LDRT and possibly other investigational treatments; and 2. A randomized, multi-institutional study for patients requiring hospitalization for hypoxemia, fever, or pulmonary compromise, but not mechanical ventilation (NCT04466683). In addition to experimental drugs, initial randomization is to LDRT 0.35 Gy or 1 Gy; after interim pre-specified evaluation, randomization will be to standard of care (SOC) or SOC plus the more efficacious/safer LDRT dose.

PRECLINICAL TRANSLATIONAL APPROACHES

While the U.S. Food and Drug Administration (FDA) has informed that radiotherapy devices are cleared for certain indications, such as tumor treatment in adults and children, their use has not been cleared for treating COVID-19 or modulating the inflammatory response in patients with COVID-19 indications. Other regulatory pathways available include emergency use authorization (EUA) ( https://bit.ly/2G5chSy) and investigational device exemptions (IDEs) ( https://bit.ly/2G9n66c). EUAs are in effect only during a public health emergency; no radiotherapy devices currently have an EUA for COVID-19. A physician may choose to use radiation therapy to treat a patient for COVID-19associated pathologies as part of the practice of medicine; however, treating clinicians should be well informed about the treatment and base their decision on solid scientific and medical evidence. IRB oversight and informed consent may be required by the physician's institution before treatment. The practice of medicine does not include use of a device in clinical trials.

The importance of considering correlative, preclinical work in the design and conduct of LDRT clinical studies was discussed using a schema of the clinical and pathological evolution of COVID-19 infection. Figure 1 shows the chronology of events that leads to the cytokine storm in SARS-CoV-2 infection, causing a dysfunctional immune response and ultimately ARDS (20).

FIG. 1

Chronology of events leading to the cytokine storm in SARS-CoV-2 infection causing dysfunctional immune response and ultimately ARDS. Published and modified, with permission, Tay MZ, Poh CM, Renia L, MacAry PA, Ng LFP. The trinity of COVID-19: immunity, inflammation and intervention. Nat Rev Immunol. 2020; 20:363–74.

The pathological evolution toward severe ailment does not always depend on viral load and culminates in a combination of symptoms. These include lymphopenia, with a relative abundance of abnormally functioning or exhausted CD8⁺ T cells, high neutrophil-to-lymphocyte ratio, increased antibody-secreting B cells, and monocytopenia with a relative increase in inflammatory CD14⁺ and CD16⁺ monocytes. Also observed are increases in multiple cytokines (e.g., IL-2, IL-6, IL-7, IL-10) and chemokines responsible for recruiting neutrophils to the lung. The evidence for some effectiveness of anti-viral therapy during the moderate phase and steroid therapy during the severe phase was also presented. Based on the discussions, suggested correlative studies for clinical trials are provided in Table 2.

TABLE 2

Suggested correlative studies in the clinic (21, 22)

A re-analysis of several animal datasets from experiments performed decades ago testing radiation treatment for bacterial and viral pneumonia reveal that results are heterogeneous, but collectively indicate lack of efficacy of radiation after infection (71). Animal models of SARS-CoV-2 infection (while not identical to human COVID-19) include non-human primates, hamsters, and transgenic mice expressing the human ACE2 receptor. However, no studies of LDRT in these models are available (3), and some workshop participants called for such preclinical experiments before continued testing of LDRT in COVID-19 patients.

RADIOBIOLOGICAL AND IMMUNOLOGICAL MECHANISMS

Using X rays to treat non-malignant disorders, including pneumonia (23), was more common in the U.S. until the 1950s, although LDRT is used in Germany, mainly for degenerative, hyperproliferative or inflammatory diseases (24), where single doses of 0.5–1.0 Gy and total doses of 3.0–6.0 Gy are delivered to patients >40 years old (24). Rapid decreases in pain and edema are commonly reported. The relevance to LDRT for COVID-19 can be questioned; however, efficacy suggests an anti-inflammatory action, supported by preclinical studies (25).

Our understanding of the biology of SARS-CoV-2 is constantly evolving and studies on the effects of radiation on COVID-19 are limited and challenging. This complex biology adds to the debate about the essential need for preclinical data and clinical trials, to include carefully specified indications, exclusions, treatment regimens, biomarkers and clinical outcomes, and the importance of a randomized study comparing radiation therapy vs. drug therapy or drug therapy ± LDRT. Selecting radiation dose may be from a randomization (noted in the study above) or by the investigators' choice.

SARS-CoV-2 in the lung targets primarily bronchial and alveolar epithelial cells, but other targets are likely, including immune cells. Pathogen-associated molecular patterns are sensed by innate pattern recognition receptors (PRR) on resident immune and non-immune cells, leading to the production of chemokines and cytokines. These molecules activate leukocyte-endothelial interactions to allow an influx of immune cells into the inflammatory site. The PRRs that sense the presence of SARS-CoV-2 have yet to be fully identified, but are likely important for the outcome of infection and could explain why this virus stimulates a vigorous hyperinflammatory response (26). Radiation is known to modulate Toll-like receptor (TLR) expression, which could alter early innate immune response (27), although it is unknown whether LDRT can do this. Immune cells, especially the myeloid subsets, are activated to phagocytose debris in the inflammatory site while dendritic cells (DC) present antigens to stimulate adaptive immunity in primary lymphoid tissues. Myeloid cells, however, appear to be dysregulated (further discussed below), and conventional DCs and pDCs are decreased in COVID-19 patients (28), but whether the generation of antiviral immunity is affected is, as yet, unclear.

Inflammation is a multistep process involving many cell types and mediators such as reactive oxygen species (ROS), nitric oxide, prostaglandins, and pro-inflammatory chemokines and cytokines, especially interleukin (IL)-1, tumor necrosis factor-α (TNF-α), IL-6, IL-8, and IL-12. Pro-inflammatory cytokines also tie into coagulation pathways and fibrin deposition, which is important in the later stages of COVID-19 pathogenesis. Vascular endothelial cells play a central role in responding to and producing cytokines that further drive cytokine production and lead to a coagulation response, mainly through the extrinsic pathway involving tissue factor and factor VIIa. Such inflammatory, pro-coagulation events can overwhelm host defenses, leading to disseminated intravascular coagulation and multiple organ failure. Coagulopathy is frequently associated with excessive systemic inflammation and is a distinct feature of advanced COVID-19 (29). The ability of single high doses of radiation to induce inflammation and trigger clotting has been known for decades (30). In contrast, little is known as to how LDRT might affect the coagulation process, particularly in ARDS. The effects of LDRT on ongoing inflammatory processes have been investigated in other model systems. However, detailed dose-response and time-course studies are essential to deciphering the pro-survival immune-modulatory response vs. cytotoxic effects of LDRT on lung vascular endothelial cells. Since endothelial cells are relatively radiosensitive, LDRT-mediated benefit may be linked to an inhibitory effect on coagulopathy, which needs to be verified. The effects of LDRT on many inflammatory processes have been investigated, but data are not consistent with respect to dose and time of response. However, a general picture has emerged of anti-inflammatory actions of doses in the range 0.1–1.0 Gy with notably reduced immune cell/endothelial cell adhesion and production of inflammatory mediators such as ROS, nitric oxide, and cytokines in vitro that translated in vivo to LDRT efficacy on experimental collagen-induced arthritis (31, 32), experimental autoimmune encephalomyelitis (33), TNF-induced polyarthritis (34), and decreased neuroinflammation in mouse models of Alzheimer's disease (35, 36). Although these LDRT studies were performed in the context of inflammatory conditions, it should be noted that LDRT was often delivered just before or after a challenge. There is a paucity of data on the effects of LDRT in preclinical animal models in the context of viral infection. Many factors, including IL-6 (37), affect radiation responses, and the use of an appropriate animal model is crucial as a preexisting inflammation can dramatically alter the response to radiation, especially if it is hyperinflammatory.

Pathophysiological changes in vessel-lung interface due to SARS-CoV-2 infection leading to the loss of vascular integrity, activation of the coagulation pathway, and inflammation have been recently reported (38) (Fig. 2). In its worst form, COVID-19 is a complex pathophysiological disease state, which includes severe pneumonia, ARDS, septic shock syndrome, disseminated intravascular coagulation, and multiorgan failure (39, 40). In patients, IL-2, IL-4, IL-6, IL-7, IL-9, IL-10, G-CSF, IP10/CXCL10, MCP1/CCL2, MIP1A, CCL3 and TNF-α are frequently elevated (41). High levels of a subset of these pro-inflammatory chemokines and cytokines, especially IL-6, correlate directly with disease severity and could be responsible for at least some of the disease's pathological features (42). Given these findings, it is not surprising that attempts are being made to target inflammation in COVID-19 (43), including IL-6 blocking antibodies, IL-1R agonists, NADPH oxidase (NOX) inhibitors, dexamethasone, JAK1 inhibitors, and now LDRT.

FIG. 2

Pathophysiological changes in vessel-lung interface due to SARS-CoV-2 infection leading to the loss of vascular integrity, activation of the coagulation pathway, and inflammation. Published and modified, with permission, Teuwen LA, Geldhof V, Pasut A, Carmeliet P. COVID-19: the vasculature unleashed. Nat Rev Immunol. 2020; 20:389–91.

Figure 3 illustrates the timeline of dynamic changes in lungs resulting from SARS-CoV-2 exposure, causing antiviral immune response and inflammatory response, leading to ARDS and multiorgan failure due to hyperinflammation. This mechanistic understanding of pathophysiological processes, resulting from early infection and pulmonary phase to the development of ARDS due to hyperinflammation, provides a “hypothetical window of opportunity” to intervene in the disease progression with LDRT; however, this concept needs to be tested and confirmed with biomarkers of immune and inflammatory responses (see Tables 2 and 3) alongside tissue oxygenation status.

FIG. 3

Timeline of the dynamic changes resulting from SARS-CoV-2 exposure. The current understanding of the pathophysiological processes from the early anti-viral immune response to the development of ARDS and multi-organ failure due to hyperinflammation, opens a “hypothetical window of opportunity” for intervention with LDRT. However, this is a concept that needs to be tested and confirmed with biomarkers of immune and inflammatory response (see Tables 2 and 3 for potential biomarkers) alongside tissue oxygenation status. The LDRT window of opportunity will likely follow the principle of as early as possible and as late as necessary but the suggestion is that LDRT will be most effective during the early stages of inflammation, with the caveat of potential adverse effects of lung irradiation and combined injury.

TABLE 3

Suggested Parameters for the Study Protocol, if LDRT is Planned to be Tested for COVID-19

TABLE 3

Continued.

TABLE 3

Continued.

In considering the cytokine profile of COVID-19, the lack of type 1 interferon (IFN) production in response to SARS-CoV-2 infection is striking. Other coronaviruses are known to interfere with IFN synthesis (44), and SARS-CoV-2 may have similar attributes. Since type 1 IFNs are important for antigen presentation, viral impairment of IFN production could have profound effects on development of adaptive anti-viral immunity and allow unrestricted viral replication. Recently, IFN-α has shown promise, either alone or with HIV-specific protease inhibitors, in treating coronavirus infections, including COVID-19 (45).

Myeloid cells appear to be dysregulated in COVID-19 and, although reports are inconsistent, increases in proinflammatory monocytes that produce high levels of IL-6 and other cytokines and have low MHC class-II expression may be related to dysfunctional innate and adaptive immunity and disease status (46). Dysfunction of alveolar macrophage by SARS-CoV-2 has also been proposed as a driver of the cytokine storm (47). Other features of the disease that correlate with its severity are T- and NK-cell depletion and exhaustion (48). Along with increases in myeloid suppressor cells (49), these outcomes contribute to a high neutrophil-to-lymphocyte ratio (41).

SARS-CoV2 infection evolves rapidly, as will the microenvironment within the tissue receiving the LDRT. Understanding the time-dose dependency of LDRT effects will be very challenging but crucial for dissecting therapeutic potential of LDRT, if any. Therefore, many questions remain about the anti-inflammatory action of LDRT in infectious diseases, and appropriate animal models of COVID-19 are urgently needed.

RISKS

Previously published studies with mouse models using combined radiation and influenza A exposure led to the hypothesis that prior lung irradiation might increase severity/susceptibility to SARS-CoV-2 and elevate the risk of pulmonary fibrosis (50). Prior SARS-CoV-2 exposure and recovery may also promote sensitivity and susceptibility to radiation-mediated lung injury and fibrosis in cancer patients. Notably, increased morbidity and mortality from SARS-CoV-2 infection was reported among lung cancer or lung metastasis patients in a multicenter study, while cancer patients without lung metastasis had no statistically significant differences compared to COVID-19 patients without cancer (51). COVID-19 patients with a history of lung radiotherapy had a poor prognosis and higher mortality risk, with a mathematical survival model having a nearly linear relationship between mortality risk after COVID-19 diagnosis and mean lung radiation dose (52).

LDRT includes exposure of whole lungs to radiation. Data from survivors of the atomic-bombings in Japan, who had average organ doses of approximately 0.2 Gy, indicate that the lifetime risk of lung cancer is higher in females than in males, the excess risk at 1 Sv for acute exposures being 2.6–6.7% vs. 1.3–2.0%, respectively (6). These ranges reflect differences in age at exposure for adults, statistical and dosimetry errors and the use of absolute vs. relative risk projection/transfer models.

Other published epidemiological studies provide similar estimates (6). The risk of fatal cancer and circulatory diseases after 1 Gy acute radiation exposure is estimated by modeling at approximately 2 to 6% (3, 53, 54), with morbidity risk about twofold higher. Smoking has a major effect on risk, as does inflammation. Overall COVID-19 mortality is <5%, with a higher loss of life expectancy than death from radiation-induced cancer or circulatory disease. However, persons above age 70 years may have a risk that is much higher than 5%. Risk estimates for adverse events from LDRT should be weighed against other treatment approaches.

OVERALL PERSPECTIVES

While there was a general consensus that LDRT for COVID-19 should be utilized only within a framework of rigorous and well-designed clinical studies, there was no agreement on whether safety data are sufficient and whether there is adequate rationale to proceed with clinical trials. Placing the role of LDRT in the context of safe and effective vaccines as well as improved anti-viral and anti-inflammatory strategies is essential.

Although the workshop did not reach consensus on whether or not clinical trials of LDRT for COVID-19 are appropriate, based on biological knowledge and clinical experience to date with the pandemic and LDRT to treat COVID-19, one major goal of the workshop was to suggest guidelines to consider for a well-designed clinical study. Taking into consideration the pros and cons of many parameters, a possible framework for a protocol was developed (Table 3). However, this framework is not a blueprint; rather, it provides guidance for those who are committed to testing LDRT as a treatment option for the COVID-19 public health crisis.

CONCLUSIONS

It is clear that more preclinical data, ideally derived from a robust model reflective of SARS-CoV-2 pathophysiology, are needed, and that little data exist regarding the pathophysiological effects of treating a lung with LDRT in the midst of a viral vasculopathy and pneumonia. However, the efforts of radiation biologists to reveal the potential mechanistic interplay between radiobiologic response and virus-induced cytokine response in a target organ was recognized. Without supporting or refuting LDRT for COVID-19 as a treatment, this review intends to assist clinicians and researchers who plan to conduct preclinical and clinical studies, given the limited evidence on the safety and efficacy of LDRT for COVID-19. Recognizing that improved treatments for severely ill patients with COVID19 are necessary, any trial that is performed must be done cautiously, with full appropriate IRB oversight and ongoing awareness of rapid changes in the biological understanding of the phases of the illness and the standard of care in this patient cohort. Also, the workshop participants recognize the potential of these novel trials to enhance understanding of the effects of radiation in modulating host immunity and anti-inflammatory response; this supports the essential inclusion of biomarkers in all studies. While the opinions of workshop participants spanned the entire arc, from “for” to “agnostic/neutral” to “against” the clinical application of LDRT, there was general agreement that collaborative preclinical research and rapid reporting of results from any clinical trials are necessary.