Time-of-day radiotherapy alters radiation pneumonitis risk but not survival in lung cancer patients with interstitial lung disease

Abstract

Background:

Lung cancer patients with comorbid interstitial lung disease (ILD) are highly susceptible to radiation pneumonitis (RP) after thoracic radiotherapy (RT). Circadian biology suggests that RT time-of-day (ToD) may influence normal-tissue injury, but evidence in ILD remains limited.

Objectives:

To evaluate the association of RT ToD with RP risk and survival outcomes.

Design:

Retrospective single-centre cohort study.

Methods:

We retrospectively analysed 424 patients with lung cancer and pre-existing ILD treated with thoracic RT. Patients were classified as morning (06:00–11:59), afternoon (12:00–17:59), or night (18:00–02:00) according to predominant treatment timing. Cox models assessed associations of ToD with RP, overall survival (OS), and progression‑free survival (PFS).

Results:

RP incidence differed by ToD: afternoon 58.3%, morning 44.1%, night 40.3% (p = 0.023), with afternoon RT showing the earliest and highest cumulative incidence. Using morning as reference, afternoon RT remained associated with higher RP risk in multivariable analysis (hazard ratio (HR) 1.42, 95% confidence interval (CI), 1.02–1.98; p = 0.036). Night RT showed no excess risk in univariable analysis, but was associated with increased RP risk after full adjustment (HR 1.62, 95% CI, 1.02–2.58; p = 0.043). In subgroup analyses using afternoon RT as reference, lower RP risk with morning and night RT was most evident in younger patients, men, and patients with NSCLC; in NSCLC, morning (HR = 0.64, 95% CI, 0.45–0.92) and night (HR = 0.46, 95% CI, 0.26–0.81) were associated with lower RP risk. RT ToD was not independently associated with OS or PFS, whereas RP was associated with worse survival.

Conclusion:

In patients with lung cancer and comorbid ILD, afternoon RT is associated with increased RP risk without an independent detriment to OS or PFS. Preferential morning scheduling, combined with careful dosimetric optimization, may help mitigate RP risk and warrants prospective validation.

Plain language summary

Radiotherapy timing affects the occurrence of radiation pneumonitis in lung cancer patients with pre-existing ILD

This study examined whether the time of day when radiotherapy is given affects the risk of radiation pneumonitis (RP) in lung cancer patients with pre-existing interstitial lung disease (ILD). The researchers reviewed records from 424 lung cancer patients with ILD who received thoracic radiotherapy. Patients were divided into three groups based on treatment timing: morning, afternoon, or night. The team then compared how often RP occurred in each group and analysed whether treatment time was related to overall survival (OS) and progression-free survival (PFS). They found that patients treated in the afternoon were more likely to develop RP than those treated in the morning or at night. However, the timing of radiotherapy itself did not clearly change OS or PFS. What mattered for survival was whether patients developed RP, as those who did generally had worse outcomes. These findings suggest that, in lung cancer patients with ILD, avoiding afternoon radiotherapy may help reduce RP risk without compromising cancer control.

Keywords

chronotherapy circadian rhythm radiation pneumonitis lung cancer interstitial lung disease

Introduction

Lung cancer remains the leading cause of cancer death worldwide and is among the most commonly diagnosed malignancies.¹ Thoracic radiotherapy (RT) serves as a pivotal treatment modality. However, the unique radiosensitivity of alveolar tissue predisposes patients to radiation-induced lung injury and progression to complications such as radiation pneumonitis (RP), thereby limiting the safe application of RT at higher doses and in broader populations.² Interstitial lung disease (ILD) encompasses a spectrum of diffuse parenchymal lung disorders characterized by inflammation and fibrosis.³ ILD is both a risk factor for the development of lung cancer and a frequent comorbidity in lung cancer populations.⁴ Previous studies have indicated that lung cancer patients with comorbid ILD have a significantly higher risk and severity of both RP and immune checkpoint inhibitor (ICI)-related pneumonitis compared to patients without ILD.⁵ Furthermore, the radiation dose to the lungs is closely associated with the occurrence of asymptomatic RP,⁶ and undergoing immunotherapy is also an independent risk factor for the development of RP in ILD patients.⁷ With the widespread application of ICIs in lung cancer treatment, the combination of RT and ICIs, while potentially producing synergistic anti-tumour effects in specific contexts, may further increase the incidence risk and severity of treatment-related pneumonitis.⁸ Therefore, this combined therapeutic strategy requires particularly careful evaluation in ILD patients whose pulmonary functional reserve is already compromised.

There is a strong biological rationale for chrono‑RT. Core circadian clock genes and downstream pathways regulate DNA‑damage recognition and repair, cell‑cycle checkpoints, and apoptosis,⁹ influence the regenerative rhythm of the pulmonary epithelium, and orchestrate diurnal immune‑cell trafficking and cytokine release,^10–12 providing a plausible basis for time‑of‑day (ToD) effects on normal‑tissue toxicity and treatment response. Clinically, studies examining ToD and RT outcomes have yielded heterogeneous results¹³: some suggest that irradiation timing modulates normal‑tissue toxicities, whereas effects on tumour control and survival are modest or inconsistent.¹⁴ Morning RT has been associated with reduced toxicity or improved outcomes in some settings, but findings vary across diseases and cohorts.^15,16 Evidence specific to thoracic toxicity – particularly RP – remains limited and conflicting,¹⁷ with some reports noting higher acute toxicity after afternoon or evening treatment and minimal survival impact,¹⁸ while others show no significant differences. Critically, it remains unknown whether RT ToD alters RP incidence or affects survival among lung cancer patients with established ILD.

To address this gap, we conducted a retrospective cohort study involving lung cancer patients with histologically confirmed disease and underlying ILD. Patients were stratified into three groups – morning, afternoon, and night¹⁵ – based on the primary treatment period during the RT course. A systematic comparison was performed to evaluate the association between RT timing and the incidence of RP, as well as survival outcomes. By assessing the impact of RT timing on RP risk in this specific high-risk population and exploring its potential implications for overall survival (OS) and progression-free survival (PFS), this study aims to provide a scientific basis for optimizing the scheduling of thoracic RT, reducing the incidence of RP, and improving patients’ quality of life.

Materials and methods

Study population

This single‑centre, retrospective cohort included consecutive patients with histologically confirmed lung cancer and pre‑existing ILD who received thoracic RT at Shandong Cancer Hospital between September 2020 and September 2024. Eligibility required: (i) newly diagnosed, pathologically verified lung cancer; (ii) radiological evidence of ILD documented before RT initiation; and (iii) availability of baseline and follow‑up thoracic imaging, baseline laboratory tests, and complete RT dosimetric data. Patients were excluded if key clinical, dosimetric, or laboratory variables were missing; if thoracic imaging at the time of radiographically suspected/confirmed pneumonitis was unavailable; or if they had received prior thoracic irradiation. The study was approved by the Ethics Committee of Shandong Cancer Hospital (SDTHEC‑202501010) and conducted in accordance with the Declaration of Helsinki and institutional guidelines.

Classification of ILD

Pre‑existing ILD was characterized by multidisciplinary review of chest CT scans (axial slice thickness 1–5 mm) and medical history obtained within 6 months before RT. Based on established criteria for incidental interstitial lung abnormalities (ILAs), CT findings were defined and classified according to Fleischner Society recommendations as non‑dependent parenchymal changes – ground‑glass or reticular opacities, architectural distortion, traction bronchiectasis, honeycombing, or non‑emphysematous cysts – involving ⩾5% of any lung zone; focal or unilateral changes, or abnormalities affecting <5% of a zone, were considered indeterminate.^19,20 For visual assessment, each lung was divided into upper, middle, and lower zones (six zones total). Using a standardized atlas, two thoracic radiologists independently scored ILA extent in every zone on a four‑point scale: 0 (0%), 1 (5% to <25%), 2 (25% to <50%), and 3 (⩾50%). Discrepancies were resolved by consensus, and interobserver agreement was recorded. The highest zonal score was taken as the overall severity grade (none = 0, mild = 1, moderate = 2, severe = 3). When fibrotic features such as traction bronchiectasis or honeycombing were present, the grade was increased by one level, up to a maximum of severe.²¹ CT patterns were then classified in line with the 2022 ATS/ERS/JRS/ALAT guideline as usual interstitial pneumonia (UIP), probable UIP, indeterminate for UIP, or alternative diagnosis.²² Because several categories contained few patients – and in keeping with prior work – we collapsed patterns into three analytic strata: UIP, probable UIP, and indeterminate/alternative²³ (Supplemental Figure 1A–C).

RT rhythm analysis

Based on prior chrono‑RT literature,^15,24 three fixed local‑time windows were prespecified for thoracic RT delivery: morning (06:00–11:59), afternoon (12:00–17:59), and night (18:00–02:00). Per‑fraction ToD was assigned by the timestamp of the first beam‑on recorded in the record‑and‑verify system (Varian ARIA; Varian Medical System, Inc., CA, USA). For multi‑field or multi‑arc fractions, the start of the first therapeutic beam determined ToD; if a fraction spanned two windows, assignment was based on the start time. For each patient, the dominant ToD for the treatment course was defined as the modal window across all thoracic fractions. In the event of a tie (equal frequency across two or more windows), the earlier window was assigned according to a prespecified hierarchy: morning > afternoon > night. All irradiation timestamps (scheduled and actual) were automatically extracted from the record‑and‑verify system; exposure classification used the actual delivery times. Quality‑assurance irradiations, simulation/verification sessions, and non‑thoracic fractions were excluded; only therapeutic thoracic fractions were included in the chronological analyses.

Treatment and assessments

Thoracic RT was delivered per institutional standards using intensity‑modulated RT (IMRT) or volumetric‑modulated arc therapy (VMAT) and stereotactic body RT (SBRT), as clinically indicated; target volumes were delineated according to departmental protocols. For each course, prescription dose and fractionation, technique, image guidance, and plan quality metrics were recorded. Thoracic dosimetry was retrospectively abstracted from the Varian Eclipse/ARIA systems and included, at minimum, mean lung dose (MLD), lung volume receiving ⩾20 Gy (V20), heart volume receiving ⩾30 Gy (V30), and lung volume receiving ⩾5 Gy (V5). Systemic therapies were administered at the treating clinician’s discretion, with start/stop dates and temporal relations to RT (neoadjuvant, concurrent, or adjuvant) extracted from the electronic medical record. Baseline clinical, radiologic, and laboratory variables were collected as detailed in the section ‘Covariates’, and disease assessments followed Response Evaluation Criteria in Solid Tumours version 1.1 (RECIST v1.1; RECIST Working Gruop) or institutional standards.

Covariates

Covariates were selected a priori using a prespecified directed acyclic graph and clinical judgement. Data were abstracted from the electronic health record and the Varian planning and record‑and‑verify systems and were cross‑validated. Patient‑level variables included age, sex, smoking history, Charlson comorbidity index (CCI), and Eastern Cooperative Oncology Group performance status (ECOG PS). ILD‑related variables comprised the high-resolution computed tomography phenotype on blinded dual read and an ILD severity score. Oncologic characteristics included primary tumour site, histology, stage, and lesion location. Systemic therapy variables captured chemotherapy regimens and timing, ICI exposure, and its temporal relationship to RT. RT variables encompassed technique, fractionation, image guidance, and thoracic dosimetric metrics. Baseline laboratory data (complete blood counts and serum albumin) were collected within 4 weeks prior to RT initiation. Inflammatory and nutritional indices derived from these laboratories – including monocyte‑to‑lymphocyte ratio (MLR), neutrophil‑to‑lymphocyte ratio (NLR), platelet‑to‑lymphocyte ratio (PLR), systemic immune‑inflammation index (SII), aggregate index of systemic inflammation (AISI), and prognostic nutritional index (PNI) – were calculated for subsequent analyses.^25–29

Follow-up and outcomes

Patients underwent daily on‑treatment reviews during RT, with the first post‑RT visit scheduled within 2–4 weeks of completion, followed by visits every 4–8 weeks through 6 months and thereafter per disease‑specific pathways; at each encounter, symptoms, physical findings, concomitant medications, and adverse events were documented, and worsening respiratory symptoms prompted unscheduled evaluation with chest imaging. The primary endpoint was RP, defined as new or progressive radiographic changes confined to or predominantly within the irradiated lung in conjunction with compatible clinical symptoms and graded per Common Terminology Criteria for Adverse Events (CTCAE v5.0; National Cancer Institute, Bethesda, MD, USA); equivocal cases were adjudicated in a blinded manner by a multidisciplinary panel (radiation oncology, radiology, pulmonology). Secondary endpoints were OS and PFS, with OS measured from RT start to death from any cause and PFS from RT start to first documented progression or death; patients without an event were censored at the date of last confirmed clinical or imaging contact, and key dates were cross‑validated against the electronic medical record and the institutional cancer registry.

Statistical analysis

Continuous variables were summarized as mean ± standard deviation or median (interquartile range (IQR)) and compared using the Student’s t test or Mann–Whitney U test, as appropriate; categorical variables were presented as number (percentage) and compared using the chi‑square test or Fisher exact test. Time‑to‑event endpoints were estimated with the Kaplan–Meier method and compared using the log‑rank test. Univariable Cox proportional hazards models were used to screen covariates; variables with p < 0.05 and those deemed clinically relevant were entered into multivariable Cox models to identify independent predictors of RP, OS, and PFS. Hazard ratios (HRs) with 95% confidence intervals (CIs) were reported. All tests were two‑sided, and p < 0.05 was considered statistically significant. Analyses were performed in R (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria). The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement for cohort studies.³⁰ The completed STROBE checklist is provided as Supplemental Table 1.

Results

Comparisons between patients with and without RP

As shown in Figure 1, a total of 424 lung cancer patients with pre-existing ILD were included in this cohort. Compared with patients without RP, those who developed RP had worse performance status (ECOG PS 2: 69.5% vs 50.0%; p < 0.001), greater comorbidity burden (high CCI: 95.5% vs 87.9%; p = 0.009), and were more likely to have type 2 diabetes (37.5% vs 7.6%; p < 0.001), a UIP pattern (23.0% vs 4.5%; p < 0.001), and severe ILD (91.0% vs 37.9%; p < 0.001). They also presented with more advanced disease (stage III–IV 87.5% vs 74.1%, p < 0.001; T3–4 50.0% vs 36.6%, p = 0.007; N1–3 87.0% vs 69.6%, p < 0.001) and were more likely to receive ICI therapy (56.0% vs 20.5%; p < 0.001). Dosimetrically, RP cases had higher thoracic exposure, including greater MLD, bilateral lung V20, heart V30, and mean heart dose (MHD; all p < 0.001). Laboratory profiles were also less favourable in the RP group, with higher inflammatory markers, including NLR, PLR, MLR, SII, and AISI, but lower PNI (all p < 0.05, Supplemental Table 2).

Figure 1.

Study overview and definition of RT ToD groups. Patients with lung cancer and concomitant ILD treated with external-beam RT at Shandong Cancer Hospital between 2020 and 2024 were identified; the primary endpoint was RP. Across the entire RT course, patients were assigned to morning (06:00–11:59), afternoon (12:00–17:59), or night (18:00–02:00) schedules as the exposure of interest; the schematic summarizes cohort assembly and group definitions.

Across ToD strata, baseline demographics, comorbidity burden, tumour location, histology, and ILD phenotype/severity were well balanced (Table 1), with a non-significant tendency towards earlier‑stage disease and more T1–2 tumours in the night cohort. In contrast, several clinical and treatment features varied by ToD. The crude incidence of RP was highest in the afternoon group (58.3%) compared with morning (44.1%) and night (40.3%; p = 0.023). ICI use was most frequent in the afternoon group (46.3%) and least common at night (21.0%; p = 0.004). RT technique distributions differed markedly (p < 0.001), with IMRT/VMAT predominant in the morning cohort (90.5%) and SBRT most often delivered at night. Night‑time RT was also characterized by fewer fractions (<28), higher dose per fraction, and higher equivalent dose in 2 Gy fractions (EQD2) and biologically effective dose (BED). Thoracic dosimetry varied with schedule: MLD was lowest in the morning and highest in the afternoon cohort, and bilateral lung V20 was higher in the afternoon group. Overall, the association between RT timing and treatment-related characteristics was stronger than its association with baseline characteristics. Therefore, its relationship with RP risk should be interpreted in conjunction with the relevant treatment features (Supplemental Figure 2).

Table 1.

Baseline and treatment characteristics by time group in lung cancer patients with comorbid ILD.

Variables	Morning (N = 254)	Afternoon (N = 108)	Night (N = 62)	p-Value
RP				0.023
No	142 (55.9%)	45 (41.7%)	37 (59.7%)
Yes	112 (44.1%)	63 (58.3%)	25 (40.3%)
Age				0.720
<65	97 (38.2%)	41 (38%)	27 (43.5%)
⩾65	157 (61.8%)	67 (62%)	35 (56.5%)
Gender				0.266
Female	52 (20.5%)	25 (23.1%)	8 (12.9%)
Male	202 (79.5%)	83 (76.9%)	54 (87.1%)
Smoke score				0.064
High	140 (55.1%)	45 (41.7%)	32 (51.6%)
Low	114 (44.9%)	63 (58.3%)	30 (48.4%)
BMI group				0.108
Normal	125 (49.2%)	51 (47.2%)	25 (40.3%)
Overweight	118 (46.5%)	47 (43.5%)	29 (46.8%)
Underweight	11 (4.3%)	10 (9.3%)	8 (12.9%)
ECOG PS				0.568
0–1	100 (39.4%)	44 (40.7%)	29 (46.8%)
2	154 (60.6%)	64 (59.3%)	33 (53.2%)
CCI				0.345
High	233 (91.7%)	101 (93.5%)	54 (87.1%)
Low	21 (8.3%)	7 (6.5%)	8 (12.9%)
T2DM				0.273
None	198 (78%)	81 (75%)	53 (85.5%)
Yes	56 (22%)	27 (25%)	9 (14.5%)
COPD				0.081
None	240 (94.5%)	106 (98.1%)	56 (90.3%)
Yes	14 (5.5%)	2 (1.9%)	6 (9.7%)
ILD				0.256
Probable UIP pattern	199 (78.3%)	83 (76.9%)	56 (90.3%)
Indeterminate and alternative diagnosis	20 (7.9%)	8 (7.4%)	2 (3.2%)
UIP pattern	35 (13.8%)	17 (15.7%)	4 (6.5%)
ILD score				0.965
Low	18 (7.1%)	7 (6.5%)	4 (6.5%)
Moderate	76 (29.9%)	31 (28.7%)	21 (33.9%)
Severe	160 (63%)	70 (64.8%)	37 (59.7%)
Stage				0.078
I–II	49 (19.3%)	16 (14.8%)	18 (29%)
III–IV	205 (80.7%)	92 (85.2%)	44 (71%)
T stage				0.089
T1–2	137 (53.9%)	62 (57.4%)	43 (69.4%)
T3–4	117 (46.1%)	46 (42.6%)	19 (30.6%)
N stage				0.218
N0	59 (23.2%)	18 (16.7%)	17 (27.4%)
N1–3	195 (76.8%)	90 (83.3%)	45 (72.6%)
M stage				0.466
M0	185 (72.8%)	76 (70.4%)	49 (79%)
M1	69 (27.2%)	32 (29.6%)	13 (21%)
Location				0.878
Central	204 (80.3%)	86 (79.6%)	48 (77.4%)
Peripheral	50 (19.7%)	22 (20.4%)	14 (22.6%)
Lung cancer subtypes				0.252
NSCLC	182 (71.7%)	85 (78.7%)	49 (79%)
SCLC	72 (28.3%)	23 (21.3%)	13 (21%)
Surgical history				0.619
None	206 (81.1%)	92 (85.2%)	50 (80.6%)
Yes	48 (18.9%)	16 (14.8%)	12 (19.4%)
Received immunotherapy				0.004
None	159 (62.6%)	58 (53.7%)	49 (79%)
Yes	95 (37.4%)	50 (46.3%)	13 (21%)
Radiation technique				<0.001
IMRT	221 (87%)	84 (77.8%)	36 (58.1%)
SBRT	24 (9.4%)	19 (17.6%)	24 (38.7%)
VMAT	9 (3.5%)	5 (4.6%)	2 (3.2%)
ChemoRT sequence				0.045
Concurrent chemoradiotherapy	82 (32.3%)	33 (30.6%)	15 (24.2%)
Sequential chemoradiotherapy	165 (65%)	71 (65.7%)	40 (64.5%)
Unknown	7 (2.8%)	4 (3.7%)	7 (11.3%)
Number of radiation sessions				0.007
<28	107 (42.1%)	51 (47.2%)	40 (64.5%)
⩾28	147 (57.9%)	57 (52.8%)	22 (35.5%)
Lung volume				0.974
High	128 (50.4%)	53 (49.1%)	31 (50%)
Low	126 (49.6%)	55 (50.9%)	31 (50%)
CTV	209.9 ± 148.5	218.6 ± 140.6	187.9 ± 173.6	0.435
CV	0.1 ± 0.0	0.1 ± 0.0	0.1 ± 0.1	0.507
Total dose	54.7 ± 8.4	56.0 ± 9.2	54.8 ± 8.9	0.393
Single dose	2.3 ± 1.3	2.7 ± 1.6	3.6 ± 2.2	<0.001
EQD2	60.8 ± 26.1	70.0 ± 30.8	85.3 ± 47.2	<0.001
BED	69.2 ± 13.5	72.6 ± 13.9	75.3 ± 18.6	0.004
MLD	8.938 ± 5.663	10.416 ± 4.200	10.048 ± 4.632	0.031
LLD	8.703 ± 7.557	9.896 ± 7.524	7.851 ± 6.750	0.189
RLD	9.974 ± 12.965	11.776 ± 8.808	11.926 ± 8.393	0.262
BV5	35.6 ± 43.3	44.1 ± 40.4	39.7 ± 15.4	0.169
BV20	14.7 ± 6.9	17.7 ± 8.5	16.1 ± 8.2	0.002
HV30	9.5 ± 14.1	11.0 ± 12.5	7.2 ± 9.5	0.192
MHD	7.728 ± 644.2	9.128 ± 7.472	7.875 ± 5.753	0.177
EDRIC	4.6 ± 2.1	5.0 ± 2.5	4.6 ± 2.4	0.263
WBC	6.1 ± 2.8	6.8 ± 4.9	6.0 ± 2.7	0.215
Monocyte	0.6 ± 0.9	0.8 ± 1.5	0.6 ± 0.3	0.098
Eosinophil	0.2 ± 1.4	0.1 ± 0.2	0.2 ± 0.2	0.902
Lymphocyte	1.4 ± 0.6	1.4 ± 0.5	1.6 ± 0.6	0.217
Neutrophil	4.0 ± 2.6	4.7 ± 4.7	3.8 ± 2.4	0.097
PLT	221.7 ± 86.4	242.1 ± 89.2	229.8 ± 91.9	0.129
Albumin	42.4 ± 4.2	42.5 ± 4.5	42.0 ± 3.4	0.731
Globulin	27.8 ± 5.3	27.5 ± 5.2	28.2 ± 5.4	0.668
NLR	3.5 ± 4.0	3.7 ± 3.4	2.8 ± 2.1	0.310
PLR	186.3 ± 120.8	193.7 ± 107.2	167.9 ± 84.8	0.352
MLR	0.5 ± 0.8	0.7 ± 1.2	0.4 ± 0.3	0.147
SII	811.7 ± 970.2	920.3 ± 1515.2	685.4 ± 504.8	0.392
PNI	49.7 ± 5.3	48.9 ± 4.7	50.5 ± 5.0	0.147
AISI	761.3 ± 2502.9	938.9 ± 3342.6	431.6 ± 507.8	0.467

AISI, aggregate index of systemic inflammation; BED, biologically effective dose; BMI, body mass index; BV5, percentage of both lungs volume receiving ⩾5 Gy; BV20, percentage of both lungs volume receiving ⩾20 Gy; CCI, Charlson comorbidity index; COPD, chronic obstructive pulmonary disease; CTV, clinical target volume; CV, clinical target volume to lung volume ratio; ECOG PS, Eastern Cooperative Oncology Group performance status; EQD2, equivalent dose in 2 Gy fractions; EDRIC, estimated dose of radiation to immune cells; HV30, percentage of heart volume receiving ⩾30 Gy; ILD, interstitial lung disease; IMRT, intensity modulated radiotherapy; LLD, left lung dose; MHD, mean heart dose; MLD, mean lung dose; MLR, monocyte to lymphocyte ratio; NLR, neutrophil to lymphocyte ratio; NSCLC, non-small cell lung cancer; PLR, platelet to lymphocyte ratio; PLT, platelet; PNI, prognostic nutritional index; RP, radiation pneumonitis; RLD, right lung dose; SBRT, stereotactic body radiotherapy; SCLC, small cell lung cancer; SII, systemic immune-inflammation index; T2DM, type 2 diabetes mellitus; UIP, usual interstitial pneumonia; VMAT, volumetric‑modulated arc therapy; WBC, white blood cell.

Association between RT ToD and RP

In time‑to‑event analyses, cumulative incidence curves of RP differed by ToD (Figure 2), with the afternoon curve rising steeply within the first month from RT start and achieving the highest plateau, whereas morning and night curves remained comparatively low throughout follow‑up. In a univariable Cox model, the HR for afternoon was 1.56 (95% CI, 0.98–2.48) and for morning was 0.99 (95% CI, 0.64–1.53). For the multivariable analyses, morning was used as the reference for interpretability. Across sequential models, afternoon RT remained associated with higher RP risk: Model 1 (unadjusted) HR = 1.57 (95% CI, 1.16–2.14; p = 0.004); Model 2 (age/gender‑adjusted) HR = 1.59 (95% CI, 1.16–2.16; p = 0.004); Model 3 (adding clinical treatment covariates) HR = 1.44 (95% CI, 1.04–2.00; p = 0.028); and Model 4 (further adjusting for thoracic dose-volume metrics/lung dose-volume metrics) HR = 1.42 (95% CI, 1.02–1.98; p = 0.036). Night RT was not different from morning in the unadjusted model (HR = 1.01; 95% CI, 0.65–1.55; p = 0.975), but after adjustment for demographics, comorbidities, ILD features, tumour stage, systemic therapy, RT technique/fractionation, and thoracic dosimetry, night was associated with higher risk (Model 3: HR = 1.63; 95% CI, 1.03–2.60; p = 0.039; Model 4: HR = 1.62; 95% CI, 1.02–2.58; p = 0.043, Table 2). Collectively, afternoon RT carried the highest RP risk, while night RT also showed a significantly higher risk than morning after comprehensive adjustment. However, given the limited sample sizes and wide CIs in some subgroups, these subgroup analyses should be interpreted with caution.

Figure 2.

Cumulative risk of RP by ToD of RT. Curves show the cumulative RP rate for morning (green), afternoon (red), and night (blue; reference) schedules with numbers at risk displayed below; tick marks indicate censoring and the dashed vertical line marks 1 month after RT initiation. The overall p value compares the three groups (log-rank test); HRs with 95% CIs annotated on the plot were estimated from Cox proportional hazards models using the night group as the reference.

Table 2.

Association between radiotherapy ToD and the risk of RP.

Variables	Model 1			Model 2			Model 3			Model 4
Variables	HR	95% CI	p-Value	HR	95% CI	p-Value	HR	95% CI	p-Value	HR	95% CI	p-Value
Morning	Ref			Ref			Ref			Ref
Afternoon	1.57	1.16–2.14	0.004	1.59	1.16–2.16	0.004	1.44	1.04–2.00	0.028	1.42	1.02–1.98	0.036
Night	1.01	0.65–1.55	0.975	0.99	0.64–1.52	0.952	1.63	1.03–2.60	0.039	1.62	1.02–2.58	0.043

Model 1: Crude. Model 2: Adjust: age, gender. Model 3: Adjust: age, gender, smoke score, BMI group, ECOG PS, CCI, T2DM, COPD, ILD score, ILD, Stage, T stage, N stage, M stage, lung cancer subtypes, surgical history, received immunotherapy, radiation technique, lung volume, ChemoRT sequence, location. Model 4: adjust: age, gender, smoke score, BMI group, ECOG PS, CCI, T2DM, COPD, ILD score, ILD, Stage, T stage, N stage, M stage, lung cancer subtypes, surgical history, received immunotherapy, radiation technique, lung volume, ChemoRT sequence, location, CTV, BV5, HV30.

BMI, body mass index; BV5, percentage of both lungs volume receiving ⩾5 Gy; CCI, Charlson comorbidity index; CI, confidence interval; COPD, chronic obstructive pulmonary disease; CTV, clinical target volume; ECOG PS, Eastern Cooperative Oncology Group performance status; ILD, interstitial lung disease; HR, hazard ratio; HV30, percentage of heart volume receiving ⩾30 Gy; RP, radiation pneumonitis; T2DM, type 2 diabetes mellitus; ToD, time-of-day.

Subgroup findings were concordant with the primary analysis (Figure 3; Supplemental Table 5). Detailed HRs and 95% CIs for these subgroup analyses are provided in Supplemental Table 5. Using afternoon as the reference, risk reductions were most pronounced in younger patients, for whom both morning and night schedules were associated with lower hazards (HR = 0.44; 95% CI, 0.26–0.72 and HR = 0.44; 95% CI, 0.21–0.93, respectively; Figure 3(a)). Among men, afternoon remained higher risk (overall p = 0.018), with morning HR = 0.64 (95% CI, 0.46–0.90) and night HR = 0.57 (95% CI, 0.34–0.95; Figure 3(b)). In NSCLC, both morning (HR = 0.64; 95% CI, 0.45–0.92) and night (HR = 0.46; 95% CI, 0.26–0.81) were associated with lower risk relative to afternoon (overall p = 0.0075; Figure 3(c)). In older patients and in women, within‑stratum differences did not reach significance (older: overall p = 0.50; women: overall p = 0.25), though point estimates favoured morning (older: HR = 0.79; women: HR = 0.57), with night near‑neutral (older: HR = 0.83; women: HR = 1.00). In SCLC, the overall comparison was significant (p = 0.012), but pairwise estimates versus afternoon were imprecise (morning HR = 0.61; 95% CI, 0.32–1.17; night HR = 1.80; 95% CI, 0.78–4.12), reflecting small numbers. Extended subgroup analyses by clinical stage, ECOG PS, type 2 diabetes, fraction number, lung volume, and ILD pattern (Supplemental Figure 3A–F) showed a consistent direction of effect: Afternoon RT was repeatedly the highest‑risk schedule, with no statistically significant reversals of the main effect. Together with the multivariable models in Table 2, these results supported an independent association between daily RT timing and RP risk – highest for afternoon – and indicate that, after comprehensive adjustment, night also carries higher risk than morning.

Figure 3.

Subgroup analyses of RP according to ToD of radiotherapy. Cumulative RP curves with numbers at risk and within-stratum statistics are shown for age (a), gender (b), and lung cancer type (c). p Values reflect log-rank tests within each stratum; HRs (95% CIs) were derived from Cox models using the afternoon group (red) as the reference, with morning (green) and night (blue) compared accordingly.

Association between RT ToD and survival outcomes

Kaplan–Meier analyses showed no meaningful differences in OS (Supplemental Figure 4A) or PFS (Supplemental Figure 4B) among morning, afternoon, and night schedules; survival curves largely overlapped and log‑rank tests were not significant. Consistently, in univariable and multivariable Cox models (Supplemental Tables 3 and 4), ToD was not independently associated with OS or PFS. Using afternoon as the reference and adjusting for demographics, comorbidities, ILD burden, tumour stage, treatment modality, lung volume, and thoracic dosimetry, HRs for morning and night remained close to unity with 95% CIs crossing 1. In contrast, survival was chiefly driven by disease‑ and host‑related factors, including ILD severity, comorbidity burden, clinical target volume, MHD, and nutritional status. Given the observed ToD–RP association (Table 2), we further fit Cox models incorporating RP as a time‑dependent covariate. RP was independently associated with inferior OS and PFS, whereas ToD showed no residual effect on survival. Overall, although afternoon (and, after full adjustment, night) RT was linked to higher RP risk, RT timing per se did not confer a survival disadvantage; prognosis was primarily determined by underlying disease and host factors and by the occurrence of RP.

Discussion

In this retrospective cohort of lung cancer patients with pre‑existing ILD, the ToD of RT was independently associated with RP but not with OS or PFS. Afternoon RT consistently conferred the highest RP risk across sequentially adjusted models, whereas night RT showed an elevated risk only after full adjustment for clinical practice patterns and dosimetry; morning RT represented the lowest‑risk schedule. These associations were directionally consistent across prespecified and extended subgroups, supporting a robust link between daily RT timing and pulmonary toxicity in this high‑risk population. By contrast, ToD showed no independent association with OS or PFS. In multivariable models, survival was primarily driven by disease and host factors. When RP was modelled as a time‑dependent covariate, any residual ToD effect on survival disappeared, indicating that the prognostic impact of ToD is likely mediated through its influence on RP rather than tumour control.

Evidence on chronoradiotherapy and thoracic toxicity remains limited and mixed. Prior studies have emphasized mucositis or skin reactions rather than lung injury, with some suggesting greater acute toxicity in the afternoon and others finding minimal or opposite effects.^31,32 In the field of lung cancer, most studies still primarily focus on establishing risk models for RP based on dosimetric parameters.³³ However, research on the timing of RT in patients with coexisting ILD remains limited. This study found that afternoon RT was associated with the highest risk of RP, which aligns with clinical observations suggesting that normal tissues in the afternoon/evening are more susceptible to damage. In contrast, the risk associated with nighttime RT was not directly apparent in the unadjusted analysis but emerged only after multivariate adjustment. Combined with the baseline table, nighttime RT was frequently associated with SBRT, fewer fractions and higher doses per fraction, EQD2, and BED, whereas afternoon RT was correlated with the use of ICI and lung dose metrics. The correlation heatmap further suggested that stronger correlations existed primarily among fractionation, technique, and dosimetric variables themselves, rather than between ToD and any individual dose parameter. This pattern indicates that the absence of additional risk observed in the unadjusted analysis may reflect the presence of negative confounding or confounding by indication, which became more evident only after adjustment for clustered treatment characteristics. Therefore, the elevated risk associated with nighttime RT after adjustment may reflect the fact that treatment time was closely intertwined with treatment patterns, and its true association could be obscured without rigorous adjustment.¹⁸

A biologically plausible rationale links ToD to RP through circadian regulation of DNA damage sensing and repair, cell-cycle checkpoints, apoptosis, and epithelial renewal in the lung.³⁴ The pulmonary clock creates temporal ‘gates’ for injury and repair across the 24-h cycle, while circadian programmes shape diurnal immune trafficking, cytokine release, oxidative stress, and profibrotic signalling.¹⁰ Disruption of these networks can promote fibrosis and alter thresholds for radiation-induced injury.³⁵ RT-induced danger signals and local inflammation may further synergize with ICI – which is also modulated by host circadian immunity – potentially amplifying pneumonitis risk at particular times of day.³⁶ In our cohort, afternoon RT coincided with higher ICI use and higher RP incidence, which is directionally consistent with an inflammation–immunity–circadian framework, although causality cannot be established.

Subgroup analyses supported the primary signal while highlighting potential effect heterogeneity. The separation favouring morning over afternoon was most pronounced in younger patients, males, and those with NSCLC. This phenomenon is associated with reduced circadian amplitude and weakened chronoeffects in the elderly population.³⁷ Male-specific immune dynamics and a stronger pro-inflammatory response may also enhance time-dependent susceptibility.³⁸ NSCLC treatment exhibits inherent heterogeneity in fractionation regimens, irradiated lung volume, and combination with ICI,³⁹ which may amplify the interactions between treatment timing and dosimetry, rendering chronoeffects more readily observable. In older adults and women, point estimates favoured morning but did not consistently reach significance, likely reflecting fewer events and limited power. In SCLC, overall differences were suggestive, but pairwise estimates were imprecise, consistent with smaller sample size and more homogeneous chest RT regimens.⁴⁰ These results should be viewed as hypothesis‑generating.

We observed no ToD related differences in OS or PFS, consistent with prior reports indicating that timing more strongly affects normal-tissue toxicity than tumour control.⁴¹ In real-world lung cancer with ILD, survival is dominated by tumour stage, systemic therapy, and overall thoracic dose burden, rendering ToD a relatively weak survival modifier.⁴² Competing risks and limited follow‑up may further obscure small survival effects. Notably, RP itself was independently associated with worse OS and PFS,⁴³ and modelling RP as a time-dependent covariate eliminated the residual impact of ToD on survival, supporting an indirect pathway in which ToD influences prognosis primarily through RP risk.

These findings have pragmatic implications. For patients with a substantial ILD burden, prioritizing morning RT scheduling – while maintaining oncologic intent – may reduce RP incidence and improve treatment tolerance without compromising disease control. ToD can be incorporated as a low‑cost planning and workflow consideration during multidisciplinary discussions, especially when high lung doses, large target volumes, or ICI exposure are anticipated. When night treatments are unavoidable, closer surveillance and proactive monitoring of laboratory and clinical indicators of pneumonitis risk are advisable.²⁷ Optimizing the temporal coordination of ICI and RT represents a plausible strategy for future study, but ToD should not be used to justify underdosing or delaying potentially curative therapy.

Several limitations warrant consideration. This is a single-centre retrospective study, and residual confounding related to clinical workflow, patient selection, and other unmeasured factors cannot be fully excluded. In routine practice, RT timing is influenced by both scheduling logistics and clinical decision-making. We defined ToD according to the most frequent treatment time window across each patient’s treatment course. Although this approach is practical for retrospective analysis, it may not fully capture day-to-day variation in treatment timing or accurately reflect underlying biological circadian rhythms. Data on chronotype, sleep–wake patterns, and the precise timing of ICI administration were unavailable. Because all patients were treated at a single institution, centre-specific patient characteristics, treatment protocols, and scheduling practices may limit the generalisability of our findings. The predominance of male patients in the cohort may further restrict extrapolation to broader populations. In addition, several subgroup analyses were based on small sample sizes and yielded imprecise estimates; these findings should therefore be interpreted as exploratory. Prospective multicentre studies with integrated circadian assessment are needed to validate these findings and clarify the clinical relevance of ToD-guided scheduling.

Conclusion

In this retrospective cohort of lung cancer patients with pre‑existing ILD, ToD of RT was independently associated with RP risk – highest with afternoon treatment, elevated at night after comprehensive adjustment, and lowest in the morning – while showing no independent association with OS or PFS. Pneumonitis itself is associated with poorer survival, indicating that treatment timing primarily modulates normal‑tissue toxicity rather than anti-tumour efficacy. Clinically, prioritizing morning treatment and employing lung‑sparing planning may help reduce RP risk without compromising disease control. Multi-centre prospective studies are needed to validate these findings and refine chrono‑RT strategies in this high‑risk population.

Supplemental Material

sj-docx-1-tam-10.1177_17588359261446805 – Supplemental material for Time-of-day radiotherapy alters radiation pneumonitis risk but not survival in lung cancer patients with interstitial lung disease

Supplemental material, sj-docx-1-tam-10.1177_17588359261446805 for Time-of-day radiotherapy alters radiation pneumonitis risk but not survival in lung cancer patients with interstitial lung disease by Haozheng Lu, Aimin Jiang, Zhaoqi Yuan, Jinming Yu and Dawei Chen in Therapeutic Advances in Medical Oncology

Footnotes

Acknowledgements

We would like to express our heartfelt gratitude to the Shandong Cancer Hospital and Institute for providing us with their invaluable learning platform. We also extend our thanks to the patients who contributed vital information to medical research.

Declarations

ORCID iDs

Aimin Jiang

Dawei Chen

Supplemental material

Supplemental material for this article is available online.

References

Zhou

Liu

, et al. Global burden of lung cancer in 2022 and projections to 2050: incidence and mortality estimates from GLOBOCAN. Cancer Epidemiol 2024; 93: 102693.

Käsmann

Dietrich

Staab-Weijnitz

, et al. Radiation-induced lung toxicity–cellular and molecular mechanisms of pathogenesis, management, and literature review. Radiat Oncol 2020; 15: 214.

Antoniou

Margaritopoulos

Tomassetti

, et al. Interstitial lung disease. Eur Respir Rev 2014; 23: 40–54.

Frank

Dagogo-Jack

Dobre

, et al. Management of lung cancer in the patient with interstitial lung disease. Oncologist 2023; 28: 12–22.

Jiang

Zhao

, et al. The impact of pre-existing interstitial lung disease on radiation and checkpoint inhibitor pneumonitis in lung cancer patients: a systematic review and meta-analysis. Ther Adv Med Oncol 2025; 17: 17588359251338624.

Jiang

Chen

, et al. V(15(Gy)) as a predictor of asymptomatic radiation pneumonitis in patients with lung cancer: a retrospective dosimetric analysis. Precis Radiat Oncol 2025; 9: 185–191.

Jiang

Yuan

, et al. Integrated machine learning risk model for predicting radiation pneumonitis in lung cancer patients with interstitial lung disease. Ann Med 2026; 58: 2634444.

, et al. Respiratory adverse effects in patients treated with immune checkpoint inhibitors in combination with radiotherapy: a systematic review and meta-analysis. Radiat Oncol 2024; 19: 134.

Gaucher

Montellier

Sassone-Corsi

Molecular cogs: interplay between circadian clock and cell cycle. Trends Cell Biol 2018; 28: 368–379.

10.

Naik

Forrest

Paul

, et al. Circadian regulation of lung repair and regeneration. JCI Insight 2024; 9: 20240308.

11.

Cash

Sephton

Woolley

, et al. The role of the circadian clock in cancer hallmark acquisition and immune-based cancer therapeutics. J Exp Clin Cancer Res 2021; 40: 119.

12.

Andersen

PAK

Reeh

Sanders

, et al. Circadian synchronization differentially modifies cytokine-mediated transcriptomic remodeling and cell death in INS-1 cells and mouse islets. iScience 2025; 28: 112431.

13.

Chan

Rowbottom

McDonald

, et al. Does the time of radiotherapy affect treatment outcomes? A review of the literature. Clin Oncol (R Coll Radiol) 2017; 29: 231–238.

14.

Xing

Yin

, et al. Comparison of radiation esophagitis associated with daytime versus evening radiotherapy in patients with esophageal carcinoma. Radiat Oncol 2025; 20: 3.

15.

Weng

Huang

, et al. Optimizing radiotherapy timing for nasopharyngeal carcinoma: the impact of radiation scheduling on survival. JCO Precis Oncol 2025; 9: e2400603.

16.

Lan

Liu

, et al. Comparing outcomes and toxicities among patients with nasopharyngeal carcinoma treated with daytime versus evening radiotherapy: a retrospective analysis with propensity score matching. Int J Cancer 2025; 157: 345–354.

17.

Shuboni-Mulligan

Breton

Smart

, et al. Radiation chronotherapy-clinical impact of treatment time-of-day: a systematic review. J Neurooncol 2019; 145: 415–427.

18.

Hsu

Hou

Huang

, et al. Differences in toxicity and outcome associated with circadian variations between patients undergoing daytime and evening radiotherapy for prostate adenocarcinoma. Chronobiol Int 2016; 33: 210–219.

19.

Hatabu

Hunninghake

Richeldi

, et al. Interstitial lung abnormalities detected incidentally on CT: a position paper from the Fleischner Society. Lancet Respir Med 2020; 8: 726–737.

20.

Washko

Lynch

Matsuoka

, et al. Identification of early interstitial lung disease in smokers from the COPDGene Study. Acad Radiol 2010; 17: 48–53.

21.

Hata

Schiebler

Lynch

, et al. Interstitial lung abnormalities: state of the art. Radiology 2021; 301: 19–34.

22.

Raghu

Remy-Jardin

Richeldi

, et al. Idiopathic pulmonary fibrosis (an update) and progressive pulmonary fibrosis in adults: an Official ATS/ERS/JRS/ALAT Clinical Practice Guideline. Am J Respir Crit Care Med 2022; 205: e18–e47.

23.

Jaula

Mattila

Lappi-Blanco

, et al. Clinical, radiological and histopathological features of patients with familial pulmonary fibrosis. Respir Res 2024; 25: 239.

24.

Bermúdez-Guzmán

Blanco-Saborío

Ramírez-Zamora

, et al. The time for chronotherapy in radiation oncology. Front Oncol 2021; 11: 687672.

25.

Gao

, et al. The predictive value of MLR for radiation pneumonia during radiotherapy of thoracic tumor patients. Cancer Manag Res 2020; 12: 8695–8701.

26.

Lee

DS.

Clinical implications of the serum platelet-to-lymphocyte ratio in the modern radiation oncology era: research update and literature review. Radiat Oncol 2024; 19: 107.

27.

Shan

Xie

, et al. Inflammation index predicts radiation-induced lung injury and prognosis in lung tumors treated with stereotactic body radiation therapy. Jpn J Radiol 2024; 42: 102–108.

28.

, et al. Subjective Global Assessment (SGA) score could be a predictive factor for radiation pneumonitis in lung cancer patients with normal pulmonary function treated by intensity-modulated radiation therapy and concurrent chemotherapy. Clin Lung Cancer 2018; 19: e211–e217.

29.

Zinellu

Collu

Nasser

, et al. The Aggregate Index of Systemic Inflammation (AISI): a novel prognostic biomarker in idiopathic pulmonary fibrosis. J Clin Med 2021; 10 20210914.

30.

von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med 2007; 147: 573–577.

31.

Farrugia

Duncan

, et al. Daily time of radiation treatment is associated with subsequent oral mucositis severity during radiotherapy in head and neck cancer patients. Cancer Epidemiol Biomarkers Prev 2020; 29: 949–955.

32.

Shukla

Gupta

Bisht

, et al. Circadian variation in radiation-induced intestinal mucositis in patients with cervical carcinoma. Cancer 2010; 116: 2031–2035.

33.

Wang

Zhang

Yang

, et al. Multicenter development of a deep learning radiomics and dosiomics nomogram to predict radiation pneumonia risk in non-small cell lung cancer. Sci Rep 2025; 15: 17106.

34.

Sancar

Lindsey-Boltz

Kang

, et al. Circadian clock control of the cellular response to DNA damage. FEBS Lett 2010; 584: 2618–2625.

35.

Hahn

Sundar

IK.

Current perspective on the role of the circadian clock and extracellular matrix in chronic lung diseases. Int J Environ Res Public Health 2023; 20 20230130.

36.

Wang

Zeng

Gül

, et al. Circadian tumor infiltration and function of CD8(+) T cells dictate immunotherapy efficacy. Cell 2024; 187: 2690–2702.e2617.

37.

Hood

Amir

The aging clock: circadian rhythms and later life. J Clin Invest 2017; 127: 437–446.

38.

Hoffmann

Liu

Seddu

, et al. Sex hormone signaling and regulation of immune function. Immunity 2023; 56: 2472–2491.

39.

Marks

Bentzen

Deasy

, et al. Radiation dose-volume effects in the lung. Int J Radiat Oncol Biol Phys 2010; 76: S70–S76.

40.

Grønberg

Killingberg

Fløtten

, et al. High-dose versus standard-dose twice-daily thoracic radiotherapy in limited-stage SCLC: final survival data, long-term toxicity, and relapse patterns in a Randomized, Open-Label, Phase II Trial. J Thorac Oncol 2025; 20: 1108–1119.

41.

Bjarnason

Mackenzie

Nabid

, et al. Comparison of toxicity associated with early morning versus late afternoon radiotherapy in patients with head-and-neck cancer: a prospective randomized trial of the National Cancer Institute of Canada Clinical Trials Group (HN3). Int J Radiat Oncol Biol Phys 2009; 73: 166–172.

42.

Park

Lim

Kang

, et al. Predictive factors for severe radiation-induced lung injury in patients with lung cancer and coexisting interstitial lung disease. Radiother Oncol 2024; 192: 110053.

43.

Nishimura

Ono

Wakuda

, et al. Prognostic impact of pneumonitis after durvalumab therapy in patients with locally advanced non-small cell lung cancer. Invest New Drugs 2022; 40: 403–410.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.54 MB

0.00 MB