Abstract
Background
Solitary pulmonary nodules (SPNs) are commonly identified on imaging studies, and guidelines exist for follow-up. Adherence to guideline recommendations varies.
Methods
We conducted a retrospective review at a single center in a metropolitan hospital to identify characteristics of patients with lung nodules and determine the rate of follow-up of these patients. A total of 1266 patient records were reviewed, and 341 patients with solitary pulmonary nodules were identified.
Results
The presence of emphysema was associated with an increased rate of follow-up with an odds ratio of 1.9. Of the patients with SPNs, 52% do not receive appropriate follow-up. Of the nodules identified, 40% were noted to be enlarged compared to prior imaging.
Conclusions
Identification of SPNs requires a standard approach to ensure follow-up for early detection of potential malignancy. Appropriate follow-up allows improvement of the rate of early cancer detection.
Keywords: Chest imaging, Incidental pulmonary nodule, Lung cancer screening, Solitary pulmonary nodules
Background
Computed tomography (CT) scans are used widely in clinical practice. Solitary pulmonary nodules (SPNs) are identified on thoracic imaging studies both incidentally and through lung cancer screening CT. The Fleischner Society guidelines provide a framework for follow-up of these incidental SPN. Adherence to follow-up for SPNs varies widely.1 Barriers to appropriate guideline-based follow-up can include forgotten incidental findings in patients admitted for extrapulmonary complaints and patient nonadherence, among other reasons.
Current Fleischner Society guideline reports suggest that 51% of smokers over the age of 50 years have lung nodules identified on CT imaging.2 In nonsmokers, SPNs have a very low risk of malignancy, often quoted at a lower than 1% risk.2 Certain features of lung nodules raise suspicion for malignancy, including size greater than 8 mm, presence of spiculated margins, and inconsistent densities within the nodule.3 In SPNs that require follow-up, guidelines exist for appropriate follow-up intervals. In patients who are admitted, the follow-up recommendations are often described in CT imaging reports in compliance with Fleischner Society guidelines. However, whether guideline recommendations are mentioned in the radiology report and whether the report is compliant with guidelines varies on the basis of the reading radiologist.4 Assuming clear reporting of nodules and guideline-based recommendations, inpatient care must be translated to outpatient follow-up in a seamless process to ensure compliance.
At a single metropolitan hospital in New York City, we sought to determine the rate of incidental lung nodule follow-up and to identify characteristics associated with increased adherence to follow-up. We describe patient demographic information, smoking status, cancer history, radiology reports, and nodule characteristics. We also collected information about prior imaging studies when present. Using these prior images, we were able to compare nodule size and characteristics to current imaging and identify the proportion of patients with enlarging nodules, placing them at higher risk.
Methods
We performed a retrospective review of patients admitted over a 6-month period in 2019. This study was exempted from Institutional Review Board review by the Northwell Health Human Research Protection Program, and informed consent was not required (HSRD20-0265, 08/10/2020). Patients who were at least 18 years of age, who had received thoracic CT scans during an in-patient admission to an academic-affiliated community hospital in New York City, and who had at least 1 solitary pulmonary nodule present were included in this study. CT imaging modalities included CT angiogram of the chest, noncontrast CT of the chest, and CT with IV contrast of the chest, which were performed for any indication. This study was not limited to CT chest imaging performed as part of lung cancer screening.
Data related to patient demographics and characteristics, including smoking history, presence of emphysema, lung nodule characteristics, follow-up recommendations, and follow-up rates were collected. Data were collected by retrospective chart review from the electronic medical records (EMRs). Follow-up was determined by retrospective review of the EMRs as follows: when follow-up was noted to be “explicitly discussed at discharge,” the discharge physician and discharge summary presented the follow-up plan, which was concordant with the radiologist report recommendation; discharge summaries were reviewed for specific follow-up pulmonologist or primary care appointment; when patients presented for those follow-up appointments, those data were reviewed, when available. When available, follow-up appointment data were reviewed to identify outpatient follow-up, including imaging. The primary aim was to determine the rate of follow-up for identified pulmonary nodules. We analyzed characteristics of patients with lung nodules, factors associated with presence of lung nodules, factors associated with follow-up, and change in nodule size when compared to prior imaging.
Statistical Analysis
Statistical analysis was performed using SAS version 9.4. Descriptive statistics (frequency distribution for categorical variables and median, interquartile range for continuous variables) were calculated. Univariate logistic regression was used to screen variables with a P value criterion of P < .05 for entry into the model selection procedure. A logistic regression model was used to determine the factors associated with follow-up appointment being made on subjects with pulmonary nodules found after CT scan.
Results
A total of 1266 patient records were reviewed, and patient characteristics were collected (Table 1). The median age was 69 years, and median body mass index (BMI) was 26. Among these patients, 50% were White, English was the listed primary language in 91%, and 94% were insured.
Table 1.
All Patient Demographics.
| All patients (N = 1266) | With nodule (N = 341) | Without nodule (N = 925) | P value | |
|---|---|---|---|---|
| Age median (IQR) | 73 (61–83) | 67 (53–79) | <.0001 | |
| Total number | 341 | 925 | ||
| Female, N (%) | 172 (50) | 459 (50) | .7962 | |
| Total number | 341 | 925 | ||
| Race, N (%) | .0685 | |||
| Total number | 333 | 872 | ||
| White | 194 (58) | 444 (51) | ||
| Black | 69 (21) | 193 (22) | ||
| Asian | 15 (5) | 37 (4) | ||
| Other/Multiracial | 55 (17) | 198 (23) | ||
| BMI (median, IQR) | 25.7 (21.8-29.9) | 26.3 (22.5-31.2) | .0182 | |
| Total number | 332 | 774 | ||
| Language, N (%) | .5222 | |||
| Total number | 336 | 895 | ||
| English | 318 (95) | 839 (94) | ||
| Spanish | 13 (4) | 33 (4) | ||
| Other | 5 (1) | 23 (3) | ||
| With insurance, N (%) | 324 (96) | 842 (94) | .0926 | |
| Total number (*N = 1236) | 337 | 899 | ||
| Smoker† | 50% | |||
| Active smoking‡ | 21% | |||
| Emphysema‡ | 26% | |||
| Follow-up was appropriate by Fleischner guideline (Total N = 133) | 62 (47) | |||
Mann–Whitney U test on age and body mass index (BMI) between patients with and without follow-up. Chi square or Fisher's exact test on sex, race, language, insurance status, smoking status, cancer history, and emphysema history.
IQR = interquartile range.
Insurance data were missing in 30 patients.
Smoking data were available in only 330 patients.
Active smoking data were available in only 165 patients.
§The presence of emphysema was found in 341 patients.
Twenty-seven percent of total patients had a pulmonary nodule identified (Table 1). The median age was 73 years, and median BMI was 25.7. Fifty-seven percent were White, 50% were never smokers, 21% were active smokers, 26% had emphysema, and 28% had a history of cancer. In patients with pulmonary nodules, 30% had prior imaging available for comparison. In 40% of patients, the nodules were seen to be enlarged in size. Of patients with lung nodules, 52% did not have follow-up. The vast majority, 94%, (1166/1236) of patients were insured (Table 1).
Characteristics of patients with SPN with and without follow-up are shown in Table 2. In patients with follow-up, 53% were current or former smokers, 25% were active smokers, 32% had emphysema, and 27% had a history of cancer. In patients without follow-up, 48% were current or former smokers, 17% were active smokers, 20% had emphysema, and 30% had a history of cancer.
Table 2.
Patient Demographics Among Patients With Lung Nodules.
| With FU (N = 165) | Without FU (N = 176) | P value | |
|---|---|---|---|
| Age median (IQR) | 71 (61–79) | 75 (61–84) | .0385 |
| Total number | 165 | 176 | |
| Female, N (%) | 86 (52) | 86 (49) | .5477 |
| Total number | 165 | 176 | |
| Race, N (%) | .8138 | ||
| Total number | 161 | 172 | |
| White | 93 (58) | 101 (59) | |
| Black | 32 (20) | 37 (22) | |
| Asian | 9 (6) | 6 (3) | |
| Other/Multiracial | 27 (17) | 28 (16) | |
| BMI (Median, IQR) | 25.3 (21.8–30.2) | 26.0 (21.8–29.8) | .7592 |
| Total number | 159 | 173 | |
| Language, N (%) | .3993 | ||
| Total number | 162 | 174 | |
| English | 152 (94) | 166 (95) | |
| Spanish | 6 (4) | 7 (4) | |
| Other | 4 (2) | 1 (1) | |
| With insurance, N (%) | 154 (96) | 170 (97) | .6549 |
| Total number | 161 | 176 | |
| Current or former smoker, N (%) | 83 (53) | 82 (48) | .3780 |
| Total number | 158 | 172 | |
| Active smoker, N (%) | 21 (25) | 14 (17) | .1961 |
| Total number | 83 | 82 | |
| Had cancer, N (%) | 44 (27) | 52 (30) | .5547 |
| Total number | 165 | 176 | |
| Had emphysema, N (%) | 53 (32) | 35 (20) | .0099 |
| Total number | 165 | 176 |
“With FU” denotes patients with follow-up; “without FU” denotes patients without follow-up. Mann-Whitney U test on age and BMI between patients with and without follow-up. Chi-square or Fisher's exact test on sex, race, language, insurance status, smoking status, cancer history, and emphysema history.
BMI = body mass index; IQR = interquartile range.
Age, BMI, race, language, insurance, CT scan, smoking status, smoking history, presence of emphysema, and history of cancer were screened. Univariate analysis showed that emphysema was associated with the follow-up status, P = .0104, (odds ratio [OR] = 1.906, 95% confidence interval [CI]: 1.164-3.123) (Table 3). That is, the odds of those with emphysema having a follow-up made was 1.906 times higher than that of those who did not have emphysema. No other factors, including age (OR = 1.011, 95% CI 0.997-1.026), BMI (OR = 0.996, 95% CI 0.967-1.027), sex (0.878, 95% CI 0.574-1.343), race (eg, Black vs White OR = 1.065, 95% CI 0.614-1.847), primary language (eg, English vs other OR = 1.638, 95% CI 0.270-9.937), insurance status (OR = 0.776, 95% CI 0.255-2.361), smoking history (OR = 0.215, 95% CI 0.788-1.872), active smoking (OR = 1.645, 95% CI 0.770-3.514), or cancer history (OR = 0.867, 95% CI 0.540-1.392) were significantly associated with a higher odds of follow-up (Table 3).
Table 3.
Characteristics Affecting Follow-Up in Patients With Lung Nodules*.
| Variables | Odds ratio (95% CI) | P value |
|---|---|---|
| Age | 1.011 (0.997–1.026) | .1302 |
| BMI | 0.996 (0.967–1.027) | .8127 |
| Sex (female vs male) | 0.878 (0.574–1.343) | .5479 |
| Race | .9185 | |
| Black vs White | 1.065 (0.614–1.847) | |
| Asian vs White | 0.614 (0.210–1.791) | |
| Other vs White | 0.955 (0.525–1.738) | |
| Unknown vs White | 0.921 (0.224–3.788) | |
| Language | .5689 | |
| English vs unknown | 1.638 (0.270–9.937) | |
| Other vs unknown | 0.375 (0.022–6.348) | |
| Spanish vs unknown | 1.750 (0.215–14.224) | |
| Insurance status | 0.776 (0.255–2.361) | .6555 |
| Ever smoker | 0.215 (0.788–1.872) | .3782 |
| Active smoker | 1.645 (0.770–3.514) | .7985 |
| Emphysema | 1.906 (1.164–3.123) | .0104 |
| History of cancer | 0.867 (0.540–1.392) | .5555 |
BMI = body mass index.
Univariate logistic regression analysis for characteristics affecting follow-up in patients with lung nodules.
Characteristics of lung nodules in patients with and without follow-up are shown in Table 4, Table 5, respectively. Patients with missing data were excluded (n = 18 and n = 36, respectively). In both groups, most patients had 1 SPN identified (43% and 34%, respectively), while 2 or 3 SPNs were identified in 32% and 33%, respectively, and >3 nodules were identified in 25% and 33%, respectively. Most nodules were solid (79% and 82%, respectively). Most nodules were located in the right upper lobe (38% and 44%, respectively). Only 23% and 3%, respectively, were spiculated, and 23% and 28%, respectively, were associated with intrathoracic lymphadenopathy. Appropriate recommendations based on the Fleischner Society guidelines for pulmonary nodules were made in 62% of patients who did not receive follow-up and 47% of patients who received follow-up (Table 4, Table 5). Forty-four percent in the no follow-up group and 31% in the follow-up group had other pulmonary findings that required follow-up (Table 4, Table 5).
Table 4.
Characteristics of Nodules in Patients With Follow-Up*.
| Characteristics of nodules | N (%) |
|---|---|
| Number of nodules (Total N = 165) | |
| 1 | 71 (43) |
| 2 or 3 | 52 (32) |
| >3 | 42 (25) |
| Location† (Total N = 165) | |
| RUL | 63 (38) |
| RML | 37 (22) |
| RLL | 50 (30) |
| LUL | 47 (28) |
| LLL | 53 (32) |
| Nodule characterization (Total N = 165) | |
| Solid | 130 (79) |
| Part solid | 20 (12) |
| Ground glass | 15 (9) |
| With spiculation (Total N = 165) | 38 (23) |
| With intrathoracic lymphadenopathy (Total N = 165) | 38 (23) |
| Had previous imaging (Total N = 165) | 41 (25) |
| Location of recommendation for follow-up (Total N = 164) | |
| Body text | 5 (3) |
| Impression | 159 (97) |
| Follow-up plans‡ (Total N = 165) | |
| Explicitly discussed in discharge | 37 (22) |
| Pulmonary appointment | 58 (35) |
| Discussed at primary care appointment | 5 (3) |
| Repeat imaging appointment made | 9 (5) |
| Inpatient diagnostic procedure performed | 8 (5) |
| Follow-up was appropriate by Fleischner guideline (Total N = 133) | 62 (47) |
Descriptive statistics for lung nodules in patients with follow-up.
RUL = right upper lobe; RML = right middle lobe; RLL = right lower lobe; LUL = left upper lobe; LLL = left lower lobe.
18 charts were missing complete data and were excluded from some analysis.
Patients could have nodules in multiple locations.
Patients could have multiple follow-up plans.
Table 5.
Characteristics of Nodules in Patients Without Follow-Up.
| Characteristics of nodules | N (%) |
|---|---|
| Number of nodules (Total N = 176) | |
| 1 | 60 (34) |
| 2 or 3 | 58 (33) |
| >3 | 58 (33) |
| Location* (Total N = 176) | |
| RUL | 78 (44) |
| RML | 51 (29) |
| RLL | 66 (38) |
| LUL | 63 (36) |
| LLL | 54 (31) |
| Nodule characterization (Total N = 176) | |
| Solid | 144 (82) |
| Part solid | 9 (5) |
| Ground glass | 23 (13) |
| With spiculation (Total N = 176) | 5 (3) |
| With intrathoracic lymphadenopathy (Total N = 176) | 49 (28) |
| Had previous imaging (Total N = 176) | 60 (34) |
| Location of recommendation for follow-up (Total N = 31) | |
| Body text | 16 (52) |
| Impression | 15 (48) |
| Follow-up plans† (Total N = 176) | |
| Explicitly discussed in discharge | 34 (19) |
| Pulmonary appointment | 39 (22) |
| Discussed at primary care appointment | 6 (3) |
| Repeat imaging appointment made | 4 (2) |
| Inpatient diagnostic procedure performed | 1 (3) |
| Follow-up was appropriate by Fleischner guideline (Total N = 172) | 107 (62) |
Descriptive statistics for lung nodules in patients without follow-up.
RUL = right upper lobe; RML = right middle lobe; RLL = right lower lobe; LUL = left upper lobe; LLL = left lower lobe.
Patients could have nodules in multiple locations.
Patients could have multiple follow-up plans. Thirty-six charts were missing data and were excluded from some analysis.
Prior imaging was compared if available (Table 6). Thirty-four percent of patients without follow-up had prior imaging available, 30% of whom had enlarging nodules.
Table 6.
Characteristics of Nodules When Prior Imaging Was Present.
| Characteristics of Nodules | N (%) |
|---|---|
| Change in size of nodule (Total N = 101) | |
| Enlarged | 40 (40) |
| Smaller | 7 (7) |
| No change | 54 (53) |
| Received follow-up (Total N = 95) | 39 (41) |
Descriptive statistics for lung nodules when prior imaging was present.
Discussion
Almost 50% of patients identified to have incidental pulmonary nodules did not have follow-up scheduled. On the basis of our findings, demographics, smoking status, language spoken, and cancer history have no effect on whether patients who are identified to have lung nodules receive follow-up. This is somewhat surprising, as prior data has demonstrated racial disparities in incidental pulmonary nodule follow-up.5 Furthermore, only a subset of our patients (4%) were primarily Spanish speaking, a population already at risk of poor follow-up for incidental pulmonary nodules.5 Thus, our findings emphasize how vulnerable this patient population is, as their nodules may not even be captured incidentally. Patients with emphysema were almost twice as likely to receive follow-up compared to those without emphysema. Moreover, a recommended follow-up plan was noted by the radiologist in the impression of the text in only 48% of patients. This finding is in line with the existing literature, suggesting varying degrees of adherence to Fleischer Society guidelines by radiologists, which has been quoted in a range of 34.7%-60.8% management consistency rate.6
In patients with prior imaging available who did not receive follow-up, 40% of nodules had increased in size. These nodules may represent increased risk of malignancy, and early diagnosis where curative therapy can be offered is crucial. A similar cohort study of 138 patients with incidental nodules at a single center found that 39% of patients were nonadherent to follow-up, and 27% of physicians were nonadherent to guideline recommendations for incidental SPNs.7 Primary care physicians may lack resources to manage incidental SPNs. If a standardized approach and algorithm, which may include referral to pulmonology, is streamlined and followed, potentially lost to follow-up nodules can be monitored. Finally, it is noteworthy to identify that the patient population reflected in this cohort was 98% insured, bringing forth further concern that rates of surveillance and follow-up for uninsured patients are likely to be even less robust and further emphasizing the need for incidental SPN follow-up.
The results of this retrospective review are consistent with existing literature that suggests the presence of gaps in adherence to follow-up of SPNs identified on imaging.8 Some studies have suggested a more streamlined approach, which includes multidisciplinary approach to SPN management,9 although this has not been validated. Moreover, who would be included in such a multidisciplinary team may vary from region to region or institute to institute. Although most patients had mention of follow-up based on Fleischner Society guidelines recommendations in the radiologist report, true follow-up adherence did not reflect those recommendations. Future efforts to create processes at a systems level are needed to ensure appropriate follow-up and early diagnosis of potentially malignant nodules. It is important to note that both patient and clinician factors affect follow-up, and identification of missed opportunities on the clinician side provides a potential area to intervene.
Our study was a single-center retrospective study performed in an urban community hospital and may not reflect the variation in practice patterns that exist across the country. As this was a retrospective review, there was limitation in certain data points; for instance, follow-up timing was not assessed and varied widely from patient to patient. Information on insurance type was not available to provide further insight about access to care. We did not evaluate data on tissue biopsy recommendations by the radiologist, nor did we evaluate factors leading to nodule biopsy. Follow-up rates were extracted from discharge summaries and the electronic medical record. Patients were not directly contacted regarding their follow-up, so follow-up rates outside of the system may not have been captured. Therefore, it is possible that follow-up rates are underestimated. Characteristics of nodules were not always explicitly noted by reading radiologists, and a single pulmonologist evaluated the lung nodules when radiologist review was not available, which may have engendered bias. This study focused on 3 primary imaging modalities of the chest, which may capture patients with a pulmonary complaint and lead to a biased population. Moreover, other imaging modalities, such as CT coronaries or CT abdomen, which has also been shown to be related to SPN follow-up,8 were excluded from this cohort, and SPNs are commonly encountered on these imaging modalities.
Further studies should examine whether there are differences in the detection rate of lung cancer, time to detection, cancer progression, and prognosis in the groups in which guideline-adherent follow-up was obtained compared with those who were not. Additionally, future studies may correlate incidental SPN follow-up with lung cancer screening eligibility.
Conclusions
This study is consistent with prior literature suggesting gaps in adherence to follow-up of identified SPNs. It is evident that awareness of the existence of Fleischner Society guidelines on follow-up for pulmonary nodules is high. In a survey of 834 radiologists, 77.8% responded that they were aware of these guidelines, but only 58.8% worked in practices that utilized them consistently.6 This study suggests that adherence to these guidelines may be lacking despite knowledge of the recommendations, and we consider multiple areas that these gaps may occur in the clinical level during hospitalization, follow-up, outpatient clinic visit, or follow-up radiologic study (ie, Swiss cheese effect). Our study helps provide further evidence that efforts to increase adherence are still needed. Literature suggests, as does this review, that lung nodules are oftentimes identified, but patients are lost to follow-up in transitions of care. Identification of pulmonary nodules on imaging requires appropriate follow-up to improve the rate of early cancer detection.
CRediT authorship contribution statement
Zein Kattih: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Data curation, Conceptualization. Jonathan A Moore: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Conceptualization. Brandon Wilson: Writing – original draft, Methodology, Investigation, Data curation. Sravani Gajjala: Writing – original draft, Methodology, Investigation, Data curation. Jacob Schwartz: Writing – original draft, Methodology, Investigation, Data curation, Conceptualization. Josef Kushner: Writing – original draft, Investigation, Data curation. Sean Zajac: Writing – review & editing, Investigation, Data curation. Akhilesh Mahajan: Writing – original draft, Methodology, Conceptualization. Tungming Leung: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Data curation, Conceptualization. Priyanka Makkar: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Conceptualization.
Declaration of competing interest
None.
Acknowledgments
Authorship
All authors had access to the data and had a role in writing/editing the final manuscript.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References
- 1.Gould MK, Altman DE, Creekmur B, et al. Guidelines for the evaluation of pulmonary nodules detected incidentally or by screening: a survey of radiologist awareness, agreement, and adherence from the watch the spot trial. J Am Coll Radiol. 2021;18(4):545–553. doi: 10.1016/j.jacr.2020.10.003. [DOI] [PubMed] [Google Scholar]
- 2.MacMahon H, Austin JH, Gamsu G, et al. Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the Fleischner Society. Radiology. 2005;237(2):395–400. doi: 10.1148/radiol.2372041887. [DOI] [PubMed] [Google Scholar]
- 3.Nasim F, Ost DE. Management of the solitary pulmonary nodule. Curr Opin Pulm Med. 2019;25(4):344–353. doi: 10.1097/MCP.0000000000000586. [DOI] [PubMed] [Google Scholar]
- 4.Kapoor N, Lacson R, Cochon L, et al. Radiologist variation in the rates of follow-up imaging recommendations made for pulmonary nodules. J Am Coll Radiol. 2021;18:896–905. doi: 10.1016/j.jacr.2020.12.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Schut RA, Mortani Barbosa EJ., Jr. Racial/ethnic disparities in follow-up adherence for incidental pulmonary nodules: an application of a cascade-of-care framework. J Am Coll Radiol. 2020;17(11):1410–1419. doi: 10.1016/j.jacr.2020.07.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Eisenberg RL, Bankier AA, Boiselle PM. Compliance with Fleischner Society guidelines for management of small lung nodules: a survey of 834 radiologists. Radiology. 2010;255(1):218–224. doi: 10.1148/radiol.09091556. [DOI] [PubMed] [Google Scholar]
- 7.Moseson EM, Wiener RS, Golden SE, et al. Patient and clinician characteristics associated with adherence: a cohort study of veterans with incidental pulmonary nodules. Ann Am Thorac Soc. 2016;13(5):651–659. doi: 10.1513/AnnalsATS.201511-745OC. [DOI] [PubMed] [Google Scholar]
- 8.Bates R, Plooster C, Croghan I, Schroeder D, McCoy C. Incidental pulmonary nodules reported on CT abdominal imaging: frequency and factors affecting inclusion in the hospital discharge summary. J Hosp Med. 2017;12(6):454–457. doi: 10.12788/jhm.2757. [DOI] [PubMed] [Google Scholar]
- 9.Roberts TJ, Lennes IT, Hawari S, et al. Integrated, multidisciplinary management of pulmonary nodules can streamline care and improve adherence to recommendations. Oncologist. 2020;25(5):431–437. doi: 10.1634/theoncologist.2019-0519. [DOI] [PMC free article] [PubMed] [Google Scholar]