Background

Bone metastases are common in patients presenting with metastatic non–small cell lung cancer (NSCLC) and are often treated with palliative radiation therapy.¹ There exists wide practice variability in palliative treatment schedules,^2,3 with multiple randomized trials showing similar pain relief between single- and multiple-fraction courses for uncomplicated metastases.^4,5 However, single-fraction treatments are relatively uncommon in the United States,⁶ despite proven advantages with regard to patient cost and convenience.⁷

The National Quality Forum (NQF) is a not-for-profit healthcare organization that recommends consensus and evidence-based quality measures to state and federal pay-for-performance programs. In August 2012, the NQF, in collaboration with the American Society for Radiation Oncology (ASTRO), endorsed measure 1822, “External Beam Radiotherapy for Bone Metastases.”⁸ With the goal of improving quality and patient convenience while reducing practice variation and treatment costs, the measure provides the following recommended fractionation schemes for the treatment of bone metastases: 30 Gy in 10 fractions, 24 Gy in 6 fractions, 20 Gy in 5 fractions, or 8 Gy in a single fraction.⁸ Although there are exclusions for reirradiation, large-volume femoral cortical involvement, sites of surgical stabilization, cord/cauda compression, and radicular pain, the recommendation is largely agnostic of tumor site, histology, or other clinical considerations. NQF 1822 was later adopted by the Hospital Outpatient Quality Reporting (HOQR) Program as measure OP-33, and the Prospective Payment System-Exempt Cancer Hospital Quality Reporting Program was adopted as measure PCH-25, both of which are value-based programs under the Centers for Medicare & Medicaid Services.^9,10 ASTRO has since withdrawn endorsement of NQF 1822 because of concerns about the complexity of reporting.¹¹ Although the continued relevance of the quality measure is questionable in the era of dose-escalation and oligometastases, NQF 1822 remains a measure of payment determination.

Because NSCLC represents the majority of palliative bone radiation courses,³ we used a national database of patients diagnosed with metastatic NSCLC to (1) assess the frequency of quality measure–concordant radiation for bone metastases as per NQF 1822; (2) describe variation in compliance by tumor, patient, and treatment characteristics; and (3) explore and consider the value, shortcomings, and future direction of radiation treatment guideline measures.

Methods

The National Cancer Database (NCDB) was used as the primary data source and is estimated to capture approximately 70% of all newly diagnosed malignancies in the United States annually. Because the primary dataset is deidentified, this project was deemed exempt from Institutional Review Board approval.

A total of 565,347 patients with metastatic NSCLC diagnosed between 2004 and 2016 were identified in the NCDB. Of these patients, 56,113 received nonstereotactic external-beam radiation therapy for bony sites of metastatic disease as part of their initial treatment. Notably, the NCDB only captures data from the patient’s initial course of treatment after diagnosis. If multiple sites receive radiation therapy during the initial course of treatment, then data from only the most “clinically significant” target of radiation are reported, with the primary site taking precedence over metastatic sites. Patients were excluded if the radiation dose or fractionation was unknown or included a boost. Furthermore, out of concern for coding errors or irregularities in the primary database, patients were excluded if the received dose was >60 Gy, the number of delivered fractions was >30, the dose per fraction was <1 or >16 Gy, or the treatment schedule was delivered to <100 patients (Figure 1).

View Full Size

Diagram illustrating criteria for inclusion.

Citation: Journal of the National Comprehensive Cancer Network 19, 13; 10.6004/jnccn.2020.7688

Abbreviations: NCDB, National Cancer Database; NSCLC, non–small cell lung cancer; RT, radiation therapy.

Treatment courses meeting 1 of the 4 NQF 1822–approved schedules were coded as compliant. All others were considered noncompliant and were grouped as 1 to 5, 6 to 10, or ≥11 fractions and at a fractional dose of <3 or ≥3 Gy. A cutpoint at 2012 was introduced, reflecting the year of the NQF 1822 adoption. A Pearson chi-square test was used to assess differences in univariable association. A multivariable modified logistic regression model was estimated to determine the association between compliant treatment courses and patient/tumor characteristics.¹² Statistical tests were based on a 2-sided significance level with α = 0.05. Data analysis was performed using STATA/IC, version 15.1 (StataCorp LLP).

Results

A total of 42,685 courses met the criteria for inclusion, of which 60.2% were compliant according to NQF 1822. Compliance by patient and tumor characteristics is shown in Table 1. Compliance at academic facilities was higher than at community or integrated network facilities (67.1% vs 56.0% and 61.2%, respectively; P<.001) and for treatment of patients diagnosed during or after 2012 (69.1% vs 52.4% before 2012; P<.001). Compliance differed by site of treatment and increased over time (Figure 2).

View Full Size

NQF 1822 compliance between 2004 and 2016 (A) for all patients and (B) by treatment site.

Citation: Journal of the National Comprehensive Cancer Network 19, 13; 10.6004/jnccn.2020.7688

Abbreviation: NQF 1822, National Quality Forum measure 1822.

Table 1.

Patient Characteristics by NQF 1822 Compliance

View Table

On multivariate analysis, statistically significant predictors of NQF 1822 compliance included year of diagnosis during or after 2012 (relative risk [RR], 1.31), receipt of systemic therapy (RR, 1.18), and academic or integrated network facilities compared with community practice (RR, 1.19 and 1.09, respectively; Table 2). Among all treatment courses, 46.7% were to 30 Gy in 10 fractions, 9.0% were to 20 Gy in 5 fractions, 4.3% were to 8 Gy in a single fraction, and <1% were to 24 Gy in 6 fractions, which were all considered compliant; 39.8% of courses were considered noncompliant (Figure 3A). Among noncompliant courses, more than half (57.8%) were noncompliant because of extended fractionation (≥11 fractions) at a low fractional dose (<3 Gy per fraction), whereas 16.5% were short (1–5 fractions) at a high fractional dose (≥3 Gy per fraction; Figure 3B). Of noncompliant courses, extended fractionation at a low fractional dose was more common before 2012; in community practice; for treatments of the skull, head, or pelvis; and among patients who received systemic therapy, whereas short treatments at a high fractional dose were more common during and after 2012, in academic practice, and for treatments of the extremities including shoulders (Table 3).

View Full Size

Relative use of the schedules for (A) all treatments and (B) NQF 1822–noncompliant treatments.

Citation: Journal of the National Comprehensive Cancer Network 19, 13; 10.6004/jnccn.2020.7688

Abbreviations: fx, fraction; NQF 1822, National Quality Forum measure 1822.

Table 2.

Factors Associated With NQF 1822–Compliant Treatments in Multivariate Analysis

View Table

Table 3.

Relative Percentage Use of Dose/Fractionation Schemes for Noncompliant Treatment Courses

View Table

Discussion

Approximately 4 of 10 palliative radiation treatments for bone metastases in this national database study were noncompliant as per the NQF 1822 quality measure. More than half of noncompliant treatments resulted from extended fractionation, although a significant proportion also resulted from short course treatments. Compliance increased over time and was highest for treatments of the extremity and for courses performed at an academic versus a community facility.

In 2018, ASTRO leadership withdrew endorsement of NQF 1822 and urged the Centers for Medicare & Medicaid Services to remove measure OP-33 from the HOQR Program.¹¹ Citing questions around the use of radiation planning versus delivery codes, the large number of measure exclusions, the complexity of reporting concurrent palliative treatment sites, and potential differences between intended versus delivered treatment schedules, ASTRO argued that the measure had become overly burdensome to accurately report. With a significant portion of recent palliative treatments not compliant with NQF 1822, coupled with the increasing number of shorter and higher dose-per-fraction schedules, these data call into question the relevance of NQF 1822 in the current era and support ASTRO’s withdrawn endorsement.

The landscape of radiation therapy for bone metastases has shifted in the almost 10 years since NQF 1822 was initially introduced. There is renewed interest in single-fraction dose escalation with the goal of improved pain response or, in some patients, more durable local control, with mixed results in recent trials.^13–15 The line between treatment for palliative versus definitive intent may seem blurred in patients who fit the evolving definition of oligometastasis.^16,17 Indeed, numerous factors must be considered when determining an appropriate palliative treatment course—tumor histology, size, and location, nearby critical structures, overall disease burden, patient performance status, patient preference, and others. In a time of emerging data and evolving expert opinions, it is possible that NQF 1822 guidelines are unnecessarily narrow and fail to reflect the full range of acceptable treatment options. That said, with nonacademic facilities displaying less NQF 1822 compliance and higher rates of extended fractionation, our data suggest that at least a portion of the practice variation is driven by factors outside of the patient case mix. Although the use of extended fractionation accounted for more than half of noncompliant courses in this study, a large proportion of noncompliant courses also resulted from shorter fractionation schedules with variation in dose per fraction. Higher doses per fraction do not generally increase costs, so it is unclear whether avoiding this type of variation in care achieves the objective of improving value.

Notwithstanding, NQF 1822 remains a measure of payment determination (OP-33) in the HOQR Program through calendar year 2020.⁹ With the forthcoming Radiation Oncology Alternative Payment Model,¹⁸ quality metrics are becoming an increasingly important part of clinical practice and reimbursement. Measures that are simple, reflect current best practice, and allow for an appropriate degree of flexibility are likely to be of highest value.

There are several limitations to this study. Chief among them are the limitations of the primary dataset (the NCDB), including the capture of only treatments performed as a component of initial management and possible inaccuracies in the coding of radiation treatment records.¹⁹ Although efforts were made to exclude patients who received nonpalliative treatment, implausible dose fractionation, and fractional doses exceeding 16 Gy, it is possible that there could be remaining inaccuracies in the data. In addition, several exclusionary criteria from NQF 1822, such as cord compression, were not available in the primary dataset. Although this limitation may have affected rates of compliance, a sensitivity analysis excluding treatments to the spine yielded similar results and associations. Furthermore, the dataset included only patients diagnosed with a primary lung cancer. Although differences may be observed with other histologies, such as breast or prostate cancer, lung cancer represents the majority of palliative radiation courses for bone metastases.^3,6,13 Finally, although we observed an increase in rates of compliance with NQF 1822 since its introduction, some of the increase predated adoption of the measure, and it is likely that additional factors have contributed to this trend.

Conclusions

Although compliance with NQF 1822 has improved since it was first introduced nearly 10 years ago, it is an inflexible measure that may fail to capture important nuances in treatment decisions including short-course treatments. More flexible quality measures that continue to promote high-quality, cost-conscious care may be warranted as the management of bone metastases continues to evolve.