Results of a Multi-institutional Phase 2 Clinical Trial for 4DCT-ventilation Functional Avoidance Thoracic Radiation Therapy

Abstract

Purpose:

Radiation pneumonitis remains a major limitation in the radiation therapy treatment of patients with lung cancer. Functional avoidance radiation therapy uses functional imaging to reduce pulmonary toxic effects by designing radiation therapy plans that reduce doses to functional regions of the lung. Lung functional imaging has been developed that uses 4-dimensional computed tomography (4DCT) imaging to calculate 4DCT-based lung ventilation (4DCT-ventilation). A phase 2 multicenter study was initiated to evaluate 4DCT-ventilation functional avoidance radiation therapy. The study hypothesis was that functional avoidance radiation therapy could reduce the rate of grade ≥2 radiation pneumonitis to 12% compared with a 25% historical rate, with the trial being positive if ≤16.4% of patients experienced grade ≥2 pneumonitis.

Methods and Materials:

Lung cancer patients receiving curative-intent radiation therapy (prescription doses of 45–75 Gy) and chemotherapy were accrued. Patient 4DCT scans were used to generate 4DCT-ventilation images. The 4DCT-ventilation images were used to generate functional avoidance plans that reduced doses to functional portions of the lung while delivering the prescribed tumor dose. Pneumonitis was evaluated by a clinician at 3, 6, and 12 months after radiation therapy.

Results:

Sixty-seven evaluable patients were accrued between April 2015 and December 2019. The median prescription dose was 60 Gy (range, 45–66 Gy) delivered in 30 fractions (range, 15–33 fractions). The average reduction in the functional volume of lung receiving ≥20 Gy with functional avoidance was 3.5% (range, 0%–12.8%). The median follow-up was 312 days. The rate of grade ≥2 radiation pneumonitis was 10 of 67 patients (14.9%; 95% upper CI, 24.0%), meeting the phase 2 criteria.

Conclusions:

4DCT-ventilation offers an imaging modality that is convenient and provides functional imaging without an extra procedure necessary. This first report of a multicenter study of 4DCT-ventilation functional avoidance radiation therapy provided data showing that the trial met phase 2 criteria and that evaluation in a phase 3 study is warranted.

Introduction

Pulmonary toxicity, particularly radiation pneumonitis, remains a major limitation in the radiation therapy treatment of patients with lung cancer.^1–3 Traditional radiation therapy techniques focus on standard lung dose metrics (mean lung dose [MLD] for example) to mitigate the risk that patients will develop thoracic side effects.¹ Treatment planning techniques using standard dose metrics assume spatially homogeneous lung function and do not differentiate between lung function heterogeneity when assessing the dose distribution. However, approximately 70% of patients with lung cancer have spatially variant lung function owing to the tumor itself (airway obstruction, for example) or accompanying thoracic comorbidities such as emphysema.⁴ To limit pulmonary toxicity during radiation and take advantage of spatial lung function heterogeneity, functional avoidance thoracic radiation therapy has been proposed.⁵ The idea of functional avoidance is to use functional imaging to differentiate between the functional versus nonfunctional regions of the lung and to use advanced treatment-planning techniques to preferentially avoid the high-functioning portions of the lung. The hypothesis of functional avoidance is that reduced doses to functional portions of the lung will reduce the probability that patients will develop pulmonary side effects.⁶

An imaging modality that has been proposed for functional radiation therapy is ventilation imaging based on 4-dimensional (4D) computed tomography (CT) data. 4DCT images are standard 3-dimensional (3D) CT images resolved into different phases of the breathing cycle. The idea is to use 4DCT imaging along with imaging processing techniques to generate a surrogate for ventilation imaging (referred to as 4DCT-ventilation).^7,8 Most patients with lung cancer undergo 4DCT imaging as part of the standard radiation planning process.⁹ Therefore, obtaining ventilation images from the 4DCT data does not require an extra imaging procedure and reduces the imaging-related cost and dose.

4DCT-ventilation has been validated against more traditional forms of lung function imaging, including nuclear medicine planar ventilation-perfusion scans,¹⁰ single-photon emission CT imaging,^11,12 and more experimental forms of functional lung imaging.^13,14 Retrospective lung cancer treatment planning studies demonstrated that it was feasible to use 4DCT-ventilation to reduce doses to functional portions of the lung without sacrificing tumor dose or exceeding standard thoracic normal-tissue dose constraints.^15,16 Toxicity modeling studies revealed that metrics that combine both 4DCT-ventilation–based function and radiation dose were more accurate in predicting radiation pneumonitis than standard dose metrics alone.^6,17,18 Modeling studies showing that dose-function metrics (based on 4DCT-ventilation) were a better predictor of radiation pneumonitis compared with standard dose metrics gave credence to the hypothesis that incorporating 4DCT-ventilation with functional avoidance may reduce the rates of pulmonary toxicity for patients with lung cancer. Based on the modeling studies, a 2-institution, 4DCT-ventilation functional avoidance clinical trial was initiated (NCT02528942). This article reports the primary trial outcome of radiation pneumonitis. Secondary outcomes reported include nonpneumonitis thoracic clinical toxicity and functional avoidance dosimetric results.

Methods and Materials

Patients

All patients included in the study provided written informed consent to participate in an ethically and institutional review board (IRB) approved trial. The trial was open to accrual at the University of Colorado and VA Eastern Colorado Health Care System (Aurora, CO, IRB #14–1856), as well as Beaumont Health System (Royal Oak, MI, IRB# 2016–037). Inclusion criteria included the presence of pathologically confirmed lung cancer (small cell lung cancer [SCLC] or non-small cell lung cancer [NSCLC]), planned curative-intent chemotherapy, ≥18 years of age, and a course of curative-intent radiation therapy. Chemotherapy and/or immunotherapy (IO) was given per standard of care. Curative-intent radiation therapy was defined as prescription doses ranging from 45 to 75 Gy. The trial also included a 4DCT-ventilation image heterogeneity criteria, which required (1) a quantitative 15% reduction in regional lung function (determined based on previous modeling studies^19,20) present near the tumor²⁰ and (2) a qualitatively noted ventilation defect (scored as a binary yes or no by the treating radiation oncology team). The idea of the 4DCT-ventilation image heterogeneity criteria is that for patients with homogeneous lung function, there are no regions to preferentially spare, whereas a heterogeneous 4DCT-ventilation spatial distribution is more likely to be amenable to preferential functional avoidance. The quantitative, 15% regional heterogeneity criteria have been previously described^19,20 and are based on nuclear medicine concepts²¹ in which the regional function in each lung lobe is estimated using geometric approximations. Exclusion criteria included patients treated with stereotactic body radiation therapy or patients treated with a palliative course of radiation (defined as prescription doses <45 Gy). There were no limitations with respect to patient performance status or baseline pulmonary function test (PFT) values.

Functional imaging and radiation treatment planning

All patients underwent standard 4DCT imaging using a gated lung CT protocol as a standard part of the radiation treatment planning process. The 4DCT images along with image processing techniques using methods previously described^6,7,22,23 and validated^10,11 were used to generate 4DCT-ventilation images. In brief, the image processing steps to generate a 4DCT-ventilation image included contouring the lungs in the end-inhalation and end-exhalation phases of the 4DCT,²³ using deformable registration methods²⁴ to link lung voxel elements in each phase, and applying a density-change–based equation to calculate ventilation in each voxel (see equation 11 in Guerrero et al⁷). An example of a 4DCT-ventilation image for 1 of the patients on the study who presented with a ventilation defect is shown in Figure 1. A functional lung contour (referred to as “functional contour”) that represented the higher-functioning portions of the lung was then created to assist in functional radiation treatment planning by applying a function threshold of ≥15% function (determined from previous modeling work^18,20) to the 4DCT-ventilation image (Fig. 1).

Fig. 1.

An example 4-dimensional computed tomography (4DCT)–ventilation image overlaid with a standard CT image. The bright colors in the 4DCT-ventilation image represent functional portions of the lung, whereas the blue and darker tones represent ventilation defect regions. The presented patient has a central mass (outlined in red) that is occluding an airway and is causing a ventilation defect shown by the 4DCT-ventilation image in the right upper lobe. The functional lung contour (generated by applying a threshold of ≥15% function) is shown overlaid over the functional image in pink.

The radiation oncologist determined standard lung gross tumor volumes and planning target volumes (PTV).²⁵ The lungs, spinal cord, heart, and esophagus were contoured as organs at risk.²⁵ All plans were generated using rotational intensity modulated radiation therapy (IMRT) techniques. Standard-of-care thoracic treatment-planning guidelines were left up to each institution and clinician and broadly followed Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC),²⁶ Radiation Therapy Oncology Group (RTOG) 0617,²⁵ NRG LU001,²⁷ and RTOG 0538 (for SCLC) guidelines. Two treatment plans were created for each patient: first, a standard thoracic treatment plan (without incorporating any functional information), and subsequently, a functional avoidance treatment plan. Although a standard lung plan could not be used for study treatment, generating a standard lung plan provided a basis for dosimetric comparison and was useful in helping guide the treatment planner in functional planning. The standard lung plan was created first and had to be clinically approved. The standard clinical lung plan was used as a baseline to generate the functional avoidance treatment plan. The functional avoidance plan was generated by altering arc geometry to avoid the functional contour whenever appropriate and by decreasing doses to the functional contour during IMRT optimization. The treatment planners were instructed not to sacrifice target coverage and standard organ-at-risk dose constraints in favor of reducing doses to the functional contour, thus ensuring that all functional-avoidance treatment plans met standard lung cancer radiation therapy criteria. Functional contour optimization objectives were left up to the treatment planner. We report standard lung metrics including the PTV coverage, PTV hotspot, MLD, volume of lung receiving ≥20 Gy (V20), maximum spinal cord dose, mean esophagus dose, and heart V40. In functional avoidance radiation therapy, it is no longer sufficient to evaluate dose metrics alone. Metrics are needed that combine dose and function. To assess dose-function, we have reported dose metrics to the functional contour, including the mean dose to the functional contour and the fV5 (volume of the functional contour receiving ≥5 Gy), fV10, fV20, and fV30. Standard and functional dosimetry results are reported for the functional plan as well as the standard lung plan and were compared using the Wilcoxon signed rank test. The feasibility of functional avoidance has been characterized by presenting the difference in functional dose metrics between the functional avoidance plan and the standard lung plan.

All patients underwent daily cone beam CT image guidance. The decision of whether to generate an adaptive treatment plan was left up to each patient’s radiation oncologist. If an adaptive plan was to be generated, another 4DCT was acquired, and the trial dictated that the adaptive plan had to be a functional avoidance plan.

Outcome assessments

The primary outcome of the study was grade ≥2 radiation pneumonitis evaluated using Common Terminology Criteria for Adverse Events, version 4.0.²⁸ Radiation pneumonitis was assessed by a clinician using clinical presentation and radiographic findings. A diagnosis of radiation pneumonitis was generally made by first performing a clinical evaluation and confirming any suspected toxicity using diagnostic imaging. Patients were assessed for toxicity during radiation therapy, at the end of radiation therapy, and 3, 6, and 12 months after the completion of radiation therapy. The trial followed patients for up to a year, and patients were considered evaluable if they completed at least the 3-month follow-up. Patient, clinical, and treatment parameters were evaluated for factor imbalances for grade ≥2 radiation pneumonitis using the Fisher exact test for categorical variables and the Mann-Whitney test for continuous variables. Additional clinical toxicity outcomes assessed included esophagitis, dyspnea, cough, and fatigue. Toxicity rates are presented with upper 95% CI. The overall survival rate for the trial cohort is presented for the 1-year follow-up period. Although not explored in the current work, additional secondary and exploratory outcome assessments included PFTs, patient-reported outcomes, and serial imaging evaluation.

Statistics

The trial was designed as a phase 2 study with a primary endpoint of grade ≥2 radiation pneumonitis to be compared against a historical control. The historical control pneumonitis rate was based on the QUANTEC lung report,¹ a systematic meta-analysis by Palma et al,² and single-institution experiences published at the time of the study design.^29–31 Combined, these studies reported grade ≥2 radiation pneumonitis rates ranging from 18% to 37%, with the 836-patient meta-analysis reporting a symptomatic pneumonitis rate of 29.8%.² Based on the reported pneumonitis data, the historical control pneumonitis rate was taken as 25%. The hypothesis of the study, therefore, was that functional avoidance radiation therapy could reduce the crude rate of grade ≥2 radiation pneumonitis to 12% compared with a grade ≥2 radiation pneumonitis rate of 25% present with a historical control. The study was designed as a Simon 2-stage design (a futility interim analysis for 17 patients is presented by Vinogradskiy et al¹⁹) with a study power of 80% and a 1-sided test with a significance of .05. The trial results were to be considered positive if ≤11 of the 67 evaluable patients (16.4%) experienced grade ≥2 radiation pneumonitis.

Results

Patients and treatment

The trial opened to accrual on April 11, 2015, and closed to accrual on December 13, 2019. A total of 101 patients provided consent, and 67 evaluable patients were accrued. Of the 101 patients who consented, 19 did not meet inclusion criteria (including 9 [8.9%] who did not meet the image heterogeneity criteria), and 15 were lost to follow-up or withdrew. Patient, clinical, and treatment factors for the 67-patient study population are presented in Table 1. Sixty percent of the patients identified as female, the median age was 65 years, 96% of the patients identified as White, and 99% identified as non-Hispanic or non-Latino. The median Karnofsky performance status score was 90 (range, 60–100), 55% of the patients had pre-existing chronic obstructive pulmonary disease, and 93% were current or former smokers. There were no patients on the study who had pre-existing interstitial lung disease. The majority of patients (79%) were diagnosed with NSCLC, and 76% of the patients had stage 3 disease. Eleven patients (16%) underwent surgery as part of their treatment, 88% received concurrent chemotherapy, and 25% were treated with IO while they were on the study. The median prescription dose was 60 Gy (range, 45–66 Gy) delivered in 30 fractions (range, 15–33 fractions). In total, 87% of the patients were treated daily, and 13% were treated with twice-daily regimens. The median PTV size was 368 cm³ (range, 50–1124 cm³; 25th quartile, 221 cm³; 75th quartile, 534 cm³). None of the patient, clinical, or treatment factors showed imbalances with respect to predicting for pneumonitis of grade 2 or greater (all P values were >.194 using the Mann-Whitney test for continuous variables and the Fisher exact test for categorical variables).

Table 1.

Patient, clinical, and treatment parameters for the study cohort

Parameter	No. (%) or median (range)	Significance*
Patients, no.	67
Sex		0.508
Female	40 (60)
Male	27 (40)
Age,y	65 (44–86)	0.449
Race		1.000
White	64 (96)
Black	2(3)
Asian-Pacific	1(1)
Ethnicity		1.000
Not Hispanic or Latino	66 (99)
Hispanic or Latino	1(1)
KPS index	90 (60–100)	0.672
COPD		0.324
Yes	37 (55)
No	30 (45)
Smoking status		0.068
Nonsmoker	5 (7)
Current smoker	16 (24)
Former smoker	46 (69)
Type of lung cancer		1.000
NSCLC	53 (79)
SCLC	14(21)
Stage		0.786
I	2 (3)*
II	6 (9)
III	51 (76)
IV	8(12)
Surgery		0.194
Yes	11 (16)
Lobectomy	10 (15)
Pneumonectomy	1(1)
No	56 (84)
Chemotherapy		0.341
Concurrent	59 (88)
Sequential	4 (6)
Induction	4 (6)
Chemotherapy agent		0.817
Cisplatin-etoposide	24 (36)
Carboplatin-paclitaxel	24 (36)
Other	19 (28)
Immunotherapy		0.259
Yes	17(25)
No	50 (75)
Radiation prescription
Total dose, Gy	60 (45–66)	0.573
Number of fractions	30 (15–33)
Fractionation pattern		0.335
Daily	58 (87)
Twice daily	9(13)
PTV size, cm3	368 (50–1124)	0.173

Abbreviations: COPD = chronic obstructive pulmonary disease; KPS = Karnofsky Performance Status; NSCLC = non-small cell lung cancer; PTV = planning target volume; SCLC = small cell lung cancer.

^*Significance values are presented evaluating risk factors and factor imbalances for radiation pneumonitis of grade 2 or greater (calculated using the Fisher exact test for categorical variables and the Mann-Whitney test for continuous variables).

^†One patient with stage I cancer was treated for a nodal recurrence.

Standard and functional dosimetry

An example of a patient with considerable differences between the functional avoidance plan compared with a standard lung cancer plan is shown in Figure 2. Compared with a standard lung plan, the patient presented in Figure 2 had a reduction of 4.2 Gy in the mean dose to the functional contour and an absolute reduction of 12.3% in V20 to the functional contour (fV20). The dosimetric results for both the functional and standard plans are shown in Table 2. The mean PTV coverage at the prescription dose was 94.7%, and MLD was 14.2 Gy. The average mean dose to the functional contour was 13.5 Gy, and the average fV20 was 21.6%. Compared with a standard lung plan, the average reduction using functional avoidance in the mean dose to the functional contour was 1.3 Gy (range, –0.1 to 4.2 Gy), and the average reduction in fV20 was 3.5% (range, 0% to –12.8%). A histogram characterizing the reduction in the functional contour fV20 doses using functional avoidance compared with corresponding standard lung plans is presented in Appendix E1.

Fig. 2.

Representative example of a trial patient with a large difference between the functional avoidance plan and a standard lung plan. The planning computed tomography, 4-dimensional computed tomography ventilation image, planning target volume (shown in red), and dose distribution are shown. The purple arrows highlight regions where the function avoidance plan (A) was able to reduce 20 Gy (orange) and 10 Gy (cyan) regions compared with a standard lung plan (B).

Table 2.

Dosimetric results for the 67-patient functional avoidance trial cohort^*

Parameter	Functional avoidance plan, mean ± SD	Nonfunctionalplan, mean ± SD	Wilcoxon signed rank test P value
PTV coverage of Rx dose (%)	94.7 ± 3.5	95.5 ± 3.7	< .001
PTV hotspot, %	21.7 ± 10.0	21.0 ± 10.4	.054
Mean lung dose, Gy	14.2 ± 3.8	14.9 ± 3.8	< .001
Lung V20, %	24.3 ± 8.6	26.3 ± 9.0	< .001
Maximum spinal cord dose, Gy	33.5 ± 8.7	32.1 ± 9.1	< .001
Mean esophagus dose, Gy	22.0 ± 8.3	22.7 ± 8.2	< .001
Esophagus volume at Rx Dose (%)	7.9 ± 9.6	6.9 ± 9.1	.534
Heart V5, %	48.6 ± 28.5	47.7 ± 29.0	.843
Heart V40, %	5.2 ± 6.3	5.3 ± 6.0	.827
Mean heart dose	10.6 ± 6.2	10.3 ± 6.2	.774
Functional avoidance structure
Mean, Gy	13.5 ± 3.8	14.9 ± 3.8	< .001
fV5, %	67.7 ± 14.2	71.1 ± 13.0	< .001
fV10, %	42.2 ± 13.5	48.6 ± 14.5	< .001
fV20, %	21.6 ± 8.9	25.1 ± 9.4	< .001
fV30, %	12.9 ± 6.8	14.7 ± 7.5	< .001

Abbreviations: PTV = planning target volume; Rx = prescription; V5, V10, V20, V30, V40 = percentage of structure volume receiving 5 Gy, 10 Gy, 20 Gy, 30 Gy, or 40 Gy or higher.

^*Dosimetry for nonfunctional plans is shown for reference.

Primary outcome, overall survival, and pulmonary toxicity

The median follow-up for the cohort was 312 days (range, 78–427 days), with 51 of 67 patients completing the full 1-year follow-up for the study. Of the 67 patients, 55 (82.1%) were alive at the 1-year follow-up time-point. The number of patients who presented with each type and maximum grade of pulmonary toxicity is shown in Table 3. Ten patients (14.9%; 95% upper CI, 24.0%) experienced grade ≥2 radiation pneumonitis which met the phase 2 study criteria accepting the alternative hypothesis that grade ≥2 pneumonitis would be <25%. The actuarial grade ≥2 radiation pneumonitis analysis with death as a competing risk event using the Fine-Gray method is presented in Appendix E2. When study data were divided according to IO, 4 of 17 patients (23.5%) who were treated with chemoradiation and IO experienced grade ≥2 pneumonitis, compared with 6 of 50 patients (12%) who did not receive IO (P = .260 using the Fisher exact test). Other notable toxicity results included 4.5% grade ≥3 radiation pneumonitis (95% CI, 11.2%) and a 7.5% (95% CI, 15.1%) rate of grade ≥3 radiation esophagitis. There was a grade 5 hemoptysis event that the study team concluded was unlikely related to the study intervention.

Table 3.

Thoracic clinical toxicity adverse events for trial cohort^*

	Patients, no. (%)
Adverse event	Grade 0	Grade 1	Grade 2	Grade 3
Pneumonitis	42 (62.7)	15 (22.4)	7 (10.4)	3 (4.5)
Esophagitis	7 (10.4)	27 (40.3)	28 (41.8)	5 (7.5)
Dyspnea	15 (22.4)	38 (56.7)	10 (14.9)	4 (6.0)
Cough	7 (10.4)	47 (70.1)	13 (19.4)	0 (0.0)
Fatigue	5 (7.5)	48 (71.6)	12 (17.9)	2 (3.0)

^*The number and percentage of patients are reported for each adverse event and toxicity grade. The data are given for each particular grade of event rather than the number of events with a given grade or higher.

Discussion

The presented study is, to our knowledge, the first study to report on a prospective, multi-institutional clinical trial using 4DCT-ventilation functional imaging for thoracic functional-avoidance radiation therapy. 4DCT-ventilation is an imaging modality uniquely suited for radiation oncology, and 4DCT images are a standard aspect of radiation treatment planning for curative-intent radiation for patients with lung cancer. Therefore, 4DCT-ventilation provides functional information without burdening the patient with an extra imaging procedure. 4DCT-ventilation provides additional imaging advantages including improved spatial resolution (compared with nuclear medicine), bimodal information (anatomic information from the CT and functional information from the 4DCT-ventilation image), and imaging that is inherently registered to the treatment-planning CT images. Studies have retrospectively demonstrated the potential advantages of using 4DCT-ventilation for functional avoidance.^6,15,17,18 To our knowledge, this study’s data are the first to validate these modeling studies showing a reduced toxicity rate in a phase 2 study. The 14.9% rate of grade grade ≥2 radiation pneumonitis compares favorably with the rates in studies that were used to establish the historical control.^1,2,29,31

The rationale for designing a phase 2 study with a radiation pneumonitis endpoint, broad inclusion criteria, and a historical control was to complete a study with a clinically relevant end point that would be feasible to accrue to in a 5-year period in a 2-institution setting. However, the trial design was susceptible to 2 considerable, interrelated limitations: (1) suboptimal study population homogeneity (with respect to radiation pneumonitis) in the patient cohort and (2) the establishment of a clear and optimal historical control. A pertinent, more recent comparison can be made between this study’s results and the toxicity reported in RTOG 0617.^25,32 RTOG 0617 reported a 21.6% rate of grade ≥2 radiation pneumonitis (combining both acute and late events) and a 20% rate of grade ≥3 pulmonary events in the 60-Gy dose arm. When RTOG 0617 data were restricted to patients treated with IMRT, the rate of grade ≥3 radiation pneumonitis (grade ≥2 radiation pneumonitis was not explicitly reported) was reported to be 3.5%³² (which is similar to the 4.5% rate of grade ≥3 radiation pneumonitis rate reported in our functional avoidance cohort). There are several key differences between this study, the studies used to set the historical pneumonitis rate, and more recent toxicity analyses. One difference is that several of the studies used to set the historical control contained 3D conformal radiation therapy (3D-CRT) data in addition to IMRT,^1,2 which has been shown to reduce pneumonitis rates compared with 3D-CRT.^30,32 Given that the 25% grade ≥2 pneumonitis rate was chosen based on publications that included some proportion of patients treated with 3D-CRT and all patients on the current study were treated with rotational IMRT, the 25% historical pneumonitis rate may have been too high. Another clinical factor important for pneumonitis consideration is that of patients diagnosed with NSCLC versus patients diagnosed with SCLC. Twenty-one percent of our study cohort was diagnosed with SCLC. Although there were no significant differences in grade ≥2 radiation pneumonitis rates between patients with SCLC and NSCLC in this study’s cohort, previous studies have shown reduced pneumonitis rates for patients with SCLC.^33,34 Additional key differences between the functional avoidance trial and other clinical studies is that this trial did not have performance status or PFT limitations (RTOG 0617 had both performance status and PFT eligibility requirements, for example). When this trial’s data are restricted to NSCLC with Karnofsky performance status scores of ≥80 (most closely mimicking the RTOG 0617 cohort), the rate of grade ≥2 pneumonitis is 15.8%, which is similar to the overall reported rate of 14.9% for the study cohort. The functional avoidance trial included patients who had cardiothoracic surgery performed as part of their lung cancer management. There were significant (P <.01, using a t test) differences in PTV for patients in the study cohort who underwent surgery (median PTV, 244 cm³) versus the PTV for patients who did not undergo surgery (median PTV, 423 cm³), and in general, surgery has been shown to be protective against pneumonitis.³⁵ The trial inclusion criteria required that patients had to have demonstrated ventilation spatial heterogeneity. Patients who have spatial ventilation heterogeneity may have inferior baseline lung function and be more susceptible to developing pulmonary toxicity. An ideal historical control pneumonitis rate would have been assessed from lung cancer patients who met spatial heterogeneity criteria and who were treated with standard chemoradiation. An additional important characteristic is that 25% of patients on this study received treatment with IO per the results of the PACIFIC study.³ When combined with radiation, IO has been shown to increase the rate of pneumonitis.^3,36,37 When this study’s data are divided according to IO, 4 of 17 patients (23.5%) experienced grade ≥2 pneumonitis, compared with 6 of 50 patients (12%) who did not receive IO.

One patient experienced a grade 5 hemoptysis event during the study. The patient had known endobronchial tumor and CT evidence of mediastinum and vascular invasion (Appendix E3). The patient was experiencing pancytopenia with thrombocytopenia and presented to the emergency department with hemoptysis and chest pain with 1 fraction of radiation therapy remaining. Although uncommon, grade 5 hemoptysis has been reported historically,³⁸ and more recently in lung cancer clinical trials³ as well as in cases of reirradiation.³⁹ The study team reviewed both the functional avoidance plan and the standard lung plan (which was prospectively generated as part of the trial) and did not find any dosimetric differences that may have increased chances of a grade-5 hemoptysis event (presented in Appendix E3). Based on tumor-specific characteristics, chemoradiation treatment resulting in thrombocytopenia, and no significant differences between the prospectively generated functional avoidance and standard lung plans, the study team concluded that the hemoptysis event was unlikely related to the study intervention.

4DCT-ventilation functional avoidance is a technologically intensive treatment paradigm in which functional imaging and image processing become an essential aspect of the treatment planning process in busy clinical environments. Owing to the technically intensive nature of functional avoidance, pertinent secondary aims of the current study were to develop quality assurance procedures suitable for busy clinical environments and to evaluate the dose-function improvement of 4DCT-ventilation functional avoidance. The quality assurance procedures have been previously described^19,20 and include sample cases for physics and dosimetry commissioning to ensure consistent 4DCT-ventilation image generation procedures and functional avoidance treatment planning techniques, increased 4DCT imaging oversight, and specific training for physics initial plan checks. The dose-function improvement with functional avoidance was characterized by calculating the difference between dose metrics to the functional contour (Table 2) between the functional avoidance plan and what the patient would have been treated with in a standard lung planning scenario. Our study noted an average reduction of 3.5% in the V20 to the functional contour, with reductions ranging from 0% to as high as 12%. The reductions in fV20 are in line with previously reported retrospective treatment planning studies.⁴⁰ The image heterogeneity criteria were designed to select patients who could benefit dosimetrically from functional avoidance. Despite the instituted image heterogeneity criteria, there was a subset of patients for whom the functional dosimetric gain was minimal (Table 2 and Appendix E2) owing to the location of the tumor, the location of functional regions, and the highly conformal treatment plans provided by modern treatment delivery techniques. Future work using knowledge-based planning,⁴¹ proton functional avoidance,⁴² and PTV hotspot manipulation⁴³ may potentially further improve functional dosimetry.

As with any study, this 4DCT-ventilation functional avoidance trial has limitations. Patients who did not complete the 3-month follow-up were excluded from the analysis. Excluding patients who did not complete the follow-up at the 3-month time point may have influenced the pneumonitis data, because certain pneumonitis events could have been missed owing to losing patients to follow-up. Because of pragmatic considerations, no consistent treatment-planning guidelines were enforced (including not enforcing standard functional contour optimization objectives); treatment planning was at the discretion of each institution and radiation oncologist. Although a paradigm of enabling treatment-planning discretion per each physician is a more clinically realistic scenario, in a clinical trial setting, the lack of consistent lung dosimetry could have affected the homogeneity of the results. Clinical grading for pneumonitis can be variable,⁴⁴ and the interpretation of historical pneumonitis rates can be challenging. Future work examining additionally collected data including PFTs and patient-reported outcomes may help further elucidate some of the patient-based advantages of functional avoidance. Shortcomings of 4DCT-ventilation have been previously described, including imaging artifacts owing to registration inaccuracy,⁴⁵ numerical instability,⁴⁶ and nonoptimal correlation results when comparing 4DCT-ventilation with other lung function imaging modalities.¹² An additional shortcoming of 4DCT-based ventilation is that only the ventilation component is computed and no perfusion information is provided. Studies have shown that about 20% of patients with lung cancer can have differing ventilation and perfusion spatial profiles.⁴⁷ Contrast-free, 4DCT-based perfusion methods have been proposed,⁴⁸ and future work will look to incorporate perfusion, in addition to ventilation, into novel treatment-planning approaches. The treatment-planning workflow used in this trial was to generate a functional contour using the 4DCT-ventilation functional image and use the functional contour for treatment plan optimization. When a functional contour is generated, a portion of the functional heterogeneity information is lost because the functional image is turned into a binary mask. There were 2 reasons for choosing to generate a functional contour in this study. First, it was the most directly implementable method in current treatment-planning systems. The other reason was that 4DCT-ventilation image information has been shown to be most robust in regions of major ventilation defects.¹¹ By generating a functional contour from the 4DCT-ventilation image information, the binary-defect versus no-defect information is most closely represented. Other studies have developed treatment planning workflows that take into account the functional heterogeneity in the 4DCT-ventilation image (NCT02843568 and NCT02308709, for example).⁴⁹ Based on previous modeling work, both the structure-based and functional-imaging–based optimization methods provide similar prediction of radiation pneumonitis.¹⁸ Since the commencement of the study, other advances have been made that address some of the limitations of 4DCT-based functional imaging^50–53 and build on functional treatment planning techniques.^18,41,42,54 For consistency purposes, the 4DCT-ventilation imaging and treatment-planning advancements introduced after the clinical trial commenced were not incorporated into the study. The data from the current trial can be used to evaluate the novel methods and further optimize 4DCT-ventilation functional avoidance. Although the current work focused on 4DCT-based methods, it should be noted that other established and novel forms of lung function imaging have been proposed for radiation therapy, including functional avoidance using single-photon emission CT,⁴³ positron emission tomography–based functional imaging,^14,55 and magnetic resonance imaging–based perfusion.⁵⁶

Conclusion

Because 4DCTs are a standard part of the treatment planning process for patients with lung cancer, 4DCT-ventilation offers an imaging modality that is convenient and provides functional imaging without an extra imaging procedure necessary. This study reports on a multicenter, prospective study of 4DCT-ventilation functional avoidance radiation therapy. The rate of grade ≥2 radiation pneumonitis was 14.9% (10 of 67 patients). The study met phase 2 criteria demonstrating reduced pneumonitis rates and provides favorable evidence for 4DCT-ventilation functional avoidance to be investigated in a phase 3 study.