Pediatric lung lesions are a group of mostly benign pulmonary anomalies with a broad spectrum of clinical disease and histopathology. Our objective was to evaluate the characteristics of children undergoing resection of a primary lung lesion and to identify preoperative risk factors for malignancy.
A retrospective cohort study was conducted by using an operative database of 521 primary lung lesions managed at 11 children’s hospitals in the United States. Multivariable logistic regression was used to examine the relationship between preoperative characteristics and risk of malignancy, including pleuropulmonary blastoma (PPB).
None of the 344 prenatally diagnosed lesions had malignant pathology (P < .0001). Among 177 children without a history of prenatal detection, 15 (8.7%) were classified as having a malignant tumor (type 1 PPB, n = 11; other PPB, n = 3; adenocarcinoma, n = 1) at a median age of 20.7 months (interquartile range, 7.9–58.1). Malignancy was associated with the DICER1 mutation in 8 (57%) PPB cases. No malignant lesion had a systemic feeding vessel (P = .0427). The sensitivity of preoperative chest computed tomography (CT) for detecting malignant pathology was 33.3% (95% confidence interval [CI]: 15.2–58.3). Multivariable logistic regression revealed that increased suspicion of malignancy by CT and bilateral disease were significant predictors of malignant pathology (odds ratios of 42.15 [95% CI, 7.43–340.3; P < .0001] and 42.03 [95% CI, 3.51–995.6; P = .0041], respectively).
In pediatric lung masses initially diagnosed after birth, the risk of PPB approached 10%. These results strongly caution against routine nonoperative management in this patient population. DICER1 testing may be helpful given the poor sensitivity of CT for identifying malignant pathology.
The role of operative management for asymptomatic lung malformations in young children remains controversial. Little is known about the risk of pleuropulmonary blastoma (PPB), a rare malignant lung tumor that can be confused with benign lesions on the basis of diagnostic imaging.
PPB was identified in nearly 10% of lung lesions with no history of antenatal detection. These results strongly caution against routine nonoperative management. DICER1 testing may be helpful in identifying PPB cases given the poor sensitivity of computed tomography.
Primary lung lesions in children represent a rare group of predominantly benign pulmonary anomalies, including congenital pulmonary airway malformations (CPAMs), bronchogenic cysts, bronchopulmonary sequestrations (BPS), and congenital lobar emphysema (CLE).1–3 Over the past 20 years, the incidence of lung lesions has increased, with more recent studies suggesting a frequency that may be as high as 1 in 2000 children.4,5 In patients with symptomatic pulmonary disease, operative resection is unequivocally the treatment of choice.6 However, in those with asymptomatic disease, the actual risk of pneumonia and cancer remains controversial.7–10 As a result, the role of surgery in the management of incidentally detected lesions is less conclusive,11–15 and nonoperative management strategies have been espoused by some pediatric surgeons and other providers worldwide.9,16,17
In lung lesions that are not detected in utero, there is increasing evidence suggesting a remote but real risk of pleuropulmonary blastoma (PPB), an embryonal malignant lung tumor associated with the DICER1 mutation and often confused with macrocytic CPAMs.18–20 First reported in 1988,21 PPB has 3 subtypes, and evidence suggests an evolutionary progression from type I (cystic) PPB in some infants to type II (solid and/or cystic) or type III (solid) PPB by between 2 and 6 years of age.22,23 Among the studies to date that have sought to estimate the incidence of PPB in pediatric lung lesions, most have been limited by small patient numbers, single-center study design, and/or lack of maternal-fetal data.24–26
Using a multi-institutional database, our primary objective was to evaluate the clinical presentation, diagnostic accuracy, and histopathology of all children undergoing resection of a pulmonary lesion diagnosed after birth. Our group hypothesized that a focused analysis on postnatally detected lung lesions would reveal a broad spectrum of clinical presentations and pathologic diagnoses and that children with type I PPB might be misdiagnosed as having benign lung disease.
Methods
A central reliance agreement (96707) was approved by the institutional review boards of all institutions associated with the Midwest Pediatric Surgery Consortium, an 11-member group of American tertiary care children’s hospitals located in 1 of 7 contiguous states serving an estimated total population of 58 million people.27 A retrospective cohort study was conducted by using an operative database of pediatric primary lung lesions (N = 521) between 2009 and 2016 as described elsewhere.28,29 Among member hospitals, operative management of all symptomatic and asymptomatic lung lesions is the standard of care. Children were identified on the basis of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes shown in Supplemental Table 5. Current Procedural Terminology codes corresponding to removal of lung tissue (32140, 32440, 32480-4, 32505, and 32662-71) were confirmed in each patient. Data were collected and stored in Research Electronic Data Capture version 8.1.20 (Vanderbilt University, Nashville, TN). Statistical analyses were performed as appropriate by using nonparametric bivariate and multivariable logistic regression models while controlling for variables associated with outcome measures with Stata (Stata Corp, College Station, TX). Significance was defined as P < .05.
Results
Baseline Characteristics
Based on 521 patients in the entire database, 177 (34.0%) were first detected in the postnatal period. There were 82 (47.7%) right-sided and 87 (50.6%) left-sided postnatally diagnosed lung masses. Three (1.7%) children had bilateral disease. Anatomic location data are presented in Supplemental Fig 2. Postnatally diagnosed lesions were most frequently located in the left upper lobe (n = 40; 23.0%), left lower lobe (n = 36; 20.7%), and right lower lobe (n = 36; 20.7%). Thirty-two (18.4%) children had concomitant comorbidities. The most common organ systems involved were cardiac (n = 8; 4.6%), neurologic (n = 6; 3.5%), and aerodigestive (n = 4; 2.3%). Three had trisomy 21 (n = 3; 1.7%).
The baseline characteristics of primary lung lesions stratified by prenatal versus postnatal diagnosis are shown in Supplemental Table 6. There were no significant differences between children with prenatally diagnosed lesions and those with lesions initialy diagnosed postnatally on the basis of sex, race and ethnicity, insurance status, gestational age at birth, and birth weight. Apgar scores were significantly lower in postnatally diagnosed cases (1-minute Apgar score: P = .0311; 5-minute Apgar score: P = .0154), and respiratory distress at birth was significantly more common (P = .0021). There was no significant difference in lesion side when compared with prenatally diagnosed cases (P = .7791).
Preoperative Evaluation
Among children with postnatally detected lung masses, a preoperative chest computed tomography (CT) scan with intravenous contrast was performed in 163 (93.1%) children at a median age of 7.1 months (interquartile range [IQR], 0.6–58.0). Forty-four (27.2%) underwent 2 or more preoperative CT scans. Based on all original CT reports, the most common diagnoses were CPAM without a systemic feeding vessel (formerly referred to as congenital cystic adenomatoid malformation; n = 64; 39.5%), CLE (n = 34; 21.0%), and BPS (n = 25; 15.4%). A malignant lesion was suspected in 7 (4.3%) patients.
Three major age cohorts were identified on the basis of indications for operative intervention (Supplemental Fig 3). Lung resection for respiratory symptoms, including tachypnea or chronic cough, was performed in 89 (50.6%) children at a median of 4.4 months (IQR, 0.9–28.7). Another 35 (19.9%) asymptomatic children underwent prophylactic resection at a median age of 10.9 months (IQR, 7.6–151.2) for an incidental lesion identified on imaging. A third cohort with a previous history of pneumonia was documented in 54 (30.7%) at a median age of resection of 42.3 months (IQR, 11.1–86.2). Other indications for surgery were less frequent and included failure to thrive (n = 9, 5.1%; median age of 3.1 months), concern for malignancy (n = 8, 4.5%; median age of 45.4 months), and previous history of pneumothorax (n = 6, 3.4%; median age of 17.9 months).
Histopathology
Histopathology data are presented in Fig 1 A–C. The most common pathology among postnatally diagnosed lesions was CPAM without a feeding vessel in 59 (34.3%) followed by CLE in 40 (23.3%), and intralobar BPS in 26 (15.1%). In contrast to no malignancies identified in 344 prenatally diagnosed specimens (median resection age of 6.1 months [IQR, 3.8–9.1]), malignant pathology was identified in 15 (8.7%) postnatally diagnosed specimens (Table 1; P < .0001) at a median resection age of 20.7 months (IQR, 7.9–56.1). Six (54.5%) of the consortium hospitals managed at least 1 (range 1–5) malignancy. The most common malignant tumor was cystic (type 1 or type 1-regressed [1r]) PPB (n = 11; 73.3%). Two (18.2%) children with cystic PPB had bilateral disease. Most patients with PPB had a negative personal or family history for PPB-associated cancers. A DICER1 mutation was documented in 8 (57%) of the 14 patients with PPB, and genetic screening of first-degree relatives was recommended in these children positive for PPB. Genetic testing was offered to all children with PPB, but 6 patients had no documentation of a DICER1 test result in the medical record.
Pathology of pediatric lung lesions. A, Donut graph of the relative distribution of all lung lesions (N = 521) within the multi-institutional collaborative based on pathologic diagnosis (adeno, adenocarcinoma). B, Donut graph of the relative distribution of prenatally diagnosed lung lesions (n = 344) revealing the relative paucity of CLE (yellow) and no cases of malignancy. C, Donut graph of the relative distribution of postnatally diagnosed lung lesions (N = 177) depicting the relative increase in both CLE (yellow) and malignant (black) diagnoses. D, Box plots revealing the relationship between median patient age (in months) at preoperative chest CT (y-axis) and pathologic diagnosis (x-axis); whiskers are the range of minimum and maximum. E, Box plots revealing the relationship between median patient age (in months) at surgical resection (y-axis) and pathologic diagnosis (x-axis). BC, bronchogenic cyst; eBPS, extralobar bronchopulmonary sequestration; iBPS, intralobar bronchopulmonary sequestration. * P < .05 (Mann–Whitney U test).
Pathology of pediatric lung lesions. A, Donut graph of the relative distribution of all lung lesions (N = 521) within the multi-institutional collaborative based on pathologic diagnosis (adeno, adenocarcinoma). B, Donut graph of the relative distribution of prenatally diagnosed lung lesions (n = 344) revealing the relative paucity of CLE (yellow) and no cases of malignancy. C, Donut graph of the relative distribution of postnatally diagnosed lung lesions (N = 177) depicting the relative increase in both CLE (yellow) and malignant (black) diagnoses. D, Box plots revealing the relationship between median patient age (in months) at preoperative chest CT (y-axis) and pathologic diagnosis (x-axis); whiskers are the range of minimum and maximum. E, Box plots revealing the relationship between median patient age (in months) at surgical resection (y-axis) and pathologic diagnosis (x-axis). BC, bronchogenic cyst; eBPS, extralobar bronchopulmonary sequestration; iBPS, intralobar bronchopulmonary sequestration. * P < .05 (Mann–Whitney U test).
Malignancy Cases (n = 15) Among 177 Postnatally Diagnosed Lung Lesions
Case No. . | Pathology . | Lobe(s) . | Sex . | Preoperative CT Diagnosis . | Resection Age, mo . | Additional Comments . |
---|---|---|---|---|---|---|
1 | PPB, type 1 | Extralobar | Male | CPAM | 1.7 | Asymptomatic, no DICER1 testing or family history |
2 | PPB, type 1 | Left upper | Female | BC | 3.6 | Respiratory symptoms, positive DICER1 result, maternal thyroid cysts |
3 | PPB, type 1 | Left upper | Male | CPAM | 4.8 | Respiratory symptoms, positive DICER1 result, no family history |
4 | PPB, type 1 | Left upper, right middle | Male | CPAM | 7.9 | Asymptomatic, no DICER1 testing |
5 | PPB, type 1 | Extralobar | Female | Other | 8.1 | Asymptomatic, no DICER1 testing, no family history |
6 | PPB, type 1 | Bilateral upper | Male | CPAM | 10.2 | Respiratory symptoms, pneumonia, positive DICER1 result, maternal thyroid cancer |
7 | PPB, type 1r | Left lower | Male | Malignancy | 18.1 | Asymptomatic, positive DICER1 result, cystic nephroma, no family history |
8 | PPB, type 1 | Right lower | Female | CPAM | 20.7 | Respiratory symptoms, positive DICER1 result, no family history |
9 | PPB, type 1 | Left lower | Male | Malignancy | 24.3 | Pneumothorax, positive DICER1 result, no family history |
10 | PPB, type 3 | Right lower | Female | Malignancy | 34.6 | Respiratory symptoms, pneumonia, positive DICER1 result, thyroid nodules |
11 | PPB, type 3 | Left lower | Male | Malignancy | 36.2 | Respiratory symptoms, no DICER1 testing or family history |
12 | PPB, unspecified | Left lower | Female | Malignancy | 56.1 | Respiratory symptoms, pneumonia, no DICER1 testing or family history |
13 | PPB, type 1r | Right upper | Female | CLE | 117.9 | Asymptomatic, presumed DICER1 syndrome, no family history |
14 | Mucinous adenocarcinoma | Left lower | Female | Other | 178.7 | Pneumonia, no family history |
15 | PPB, type 1r | Right lower | Female | Other | 187.8 | Pneumothorax, no DICER1 testing, maternal leukemia |
Case No. . | Pathology . | Lobe(s) . | Sex . | Preoperative CT Diagnosis . | Resection Age, mo . | Additional Comments . |
---|---|---|---|---|---|---|
1 | PPB, type 1 | Extralobar | Male | CPAM | 1.7 | Asymptomatic, no DICER1 testing or family history |
2 | PPB, type 1 | Left upper | Female | BC | 3.6 | Respiratory symptoms, positive DICER1 result, maternal thyroid cysts |
3 | PPB, type 1 | Left upper | Male | CPAM | 4.8 | Respiratory symptoms, positive DICER1 result, no family history |
4 | PPB, type 1 | Left upper, right middle | Male | CPAM | 7.9 | Asymptomatic, no DICER1 testing |
5 | PPB, type 1 | Extralobar | Female | Other | 8.1 | Asymptomatic, no DICER1 testing, no family history |
6 | PPB, type 1 | Bilateral upper | Male | CPAM | 10.2 | Respiratory symptoms, pneumonia, positive DICER1 result, maternal thyroid cancer |
7 | PPB, type 1r | Left lower | Male | Malignancy | 18.1 | Asymptomatic, positive DICER1 result, cystic nephroma, no family history |
8 | PPB, type 1 | Right lower | Female | CPAM | 20.7 | Respiratory symptoms, positive DICER1 result, no family history |
9 | PPB, type 1 | Left lower | Male | Malignancy | 24.3 | Pneumothorax, positive DICER1 result, no family history |
10 | PPB, type 3 | Right lower | Female | Malignancy | 34.6 | Respiratory symptoms, pneumonia, positive DICER1 result, thyroid nodules |
11 | PPB, type 3 | Left lower | Male | Malignancy | 36.2 | Respiratory symptoms, no DICER1 testing or family history |
12 | PPB, unspecified | Left lower | Female | Malignancy | 56.1 | Respiratory symptoms, pneumonia, no DICER1 testing or family history |
13 | PPB, type 1r | Right upper | Female | CLE | 117.9 | Asymptomatic, presumed DICER1 syndrome, no family history |
14 | Mucinous adenocarcinoma | Left lower | Female | Other | 178.7 | Pneumonia, no family history |
15 | PPB, type 1r | Right lower | Female | Other | 187.8 | Pneumothorax, no DICER1 testing, maternal leukemia |
Perioperative Data
The median weight and age at surgical resection were 9.4 kg (IQR, 5.2–20.0) and 11.1 months (IQR, 3.1–72.0), respectively. Twenty-four (13.6%) underwent lung resection during the newborn period (median, 2.1 weeks; IQR, 1.5–2.9), whereas 66 (37.5%) had resections performed between 1 and 12 months (median, 5.7 months; IQR, 2.5–8.6). Eighty-six (48.9%) children had surgery after 12 months of age (median, 75.3 months; IQR, 34.5–158.4).
Data on the relationship between pathologic diagnosis and patient age are shown in Fig 1 D and E. Children with CLE were significantly younger (median 1.5 months at resection [IQR, 0.6–5.6]) compared with those with other types of pathology (P < .001). By comparison, the median age at resection in the malignancy and CPAM cohorts was similar: 20.7 months (IQR, 7.9–56.1) and 17.9 months (IQR, 5.1–83.2), respectively (P > .9999). Although not statistically significant, cystic PPB was associated with younger aged patients (median, 10.2 months; IQR, 4.8–24.4) when compared with those with other malignant diagnoses (median, 46.2 months; IQR, 35.0–148.1; P = .0557).
The indications for operative management based on a given pathologic diagnosis are shown in Supplemental Table 7. Intralobar BPS lesions were associated with a relative high rate of previous pneumonia (69%), whereas most CLE lesions were associated with respiratory symptoms and failure to thrive compared with other pathologies. Compared with CPAM, malignant lesions had higher rates of pneumothoraces, but this was not statistically significant (P = .1026).
Children undergoing thoracotomy procedures were significantly younger than those undergoing thoracoscopic resection (median, 6.1 vs 35.0 months, respectively; P = .0308). Twelve (80%) children with malignant lesions underwent resection by thoracotomy, whereas 2 (13.3%) had a thoracoscopic approach (Supplemental Table 8). There was 1 (6.7%) case of thoracoscopy converted to thoracotomy. For the entire postnatally diagnosed lung lesion cohort, there were 123 (70.7%) lobectomies, 30 (17.2%) simple excisions (eg, for extralobar BPS), and 15 (8.6%) nonanatomic resections. All 3 patients with bilateral disease underwent an additional resection procedure on the contralateral side.
For the entire postnatally diagnosed cohort, the median chest tube duration was 3.0 days (IQR, 2.0–5.0), and the median postoperative hospital length of stay was 4.0 days (IQR, 3.0–8.0). There was no significant difference in hospital length of stay between children with benign disease (median, 5.0 days; IQR [3.0–8.0]) and those with malignant lesions (median, 4.0 days; IQR [3.0–6.0]; P = .2534). Two (1.1%) children postnatally diagnosed with lung lesions were placed on extracorporeal membrane oxygenation support in the postoperative period. A total of 173 (98.9%) children survived to hospital discharge. At least 1 postoperative 30-day complication occurred in 45 (26.0%) patients. The most common complications were intubation >48 hours (n = 14; 8.1%) and pneumothorax requiring insertion of a new chest tube (n = 13; 7.5%).
Radiology-Pathology Correlation
The overall diagnostic accuracy rate of CT for determining the correct lung pathology was 76.3%. Sensitivity, specificity, and predictive value data on preoperative chest CT as a predictor of histopathology are shown in Table 2. CLE and CPAM were associated with the highest sensitivity (77.5% [95% confidence interval (CI), 62.5–87.7] and 77.0% [95% CI, 65.1–85.8], respectively). A suspected malignancy was documented by preoperative CT (mean, 1.6 ± 1.2 scans) in one-third (n = 5, 33.3%) of malignancy cases. Five (33.3%) cystic PPB lesions were interpreted by the radiologist as CPAMs.
Diagnostic Accuracy of Preoperative Chest CT in Postnatally Diagnosed Lung Lesions (n = 163)
Pathology . | Sensitivity . | Specificity . | Positive Predictive Value . | Negative Predictive Value . |
---|---|---|---|---|
CPAM | 77.0 (65.1–85.8) | 83.6 (75.8–89.3) | 71.2 (59.4–80.7) | 87.4 (79.9–92.3) |
BPS | 65.8 (49.9–78.8) | 97.8 (93.9–99.4) | 89.3 (72.8–96.3) | 91.3 (85.7–94.8) |
CLE | 77.5 (62.5–87.7) | 97.8 (93.8–99.4) | 91.2 (77.0–97.0) | 93.7 (88.5–96.7) |
Bronchogenic cyst | 50.0 (31.4–68.6) | 99.4 (96.4–100) | 92.3 (66.7–99.6) | 92.7 (87.7–95.8) |
Malignancy | 33.3 (15.2–58.3) | 98.8 (95.6–99.8) | 71.4 (35.9–94.9) | 94.1 (89.5–96.8) |
Pathology . | Sensitivity . | Specificity . | Positive Predictive Value . | Negative Predictive Value . |
---|---|---|---|---|
CPAM | 77.0 (65.1–85.8) | 83.6 (75.8–89.3) | 71.2 (59.4–80.7) | 87.4 (79.9–92.3) |
BPS | 65.8 (49.9–78.8) | 97.8 (93.9–99.4) | 89.3 (72.8–96.3) | 91.3 (85.7–94.8) |
CLE | 77.5 (62.5–87.7) | 97.8 (93.8–99.4) | 91.2 (77.0–97.0) | 93.7 (88.5–96.7) |
Bronchogenic cyst | 50.0 (31.4–68.6) | 99.4 (96.4–100) | 92.3 (66.7–99.6) | 92.7 (87.7–95.8) |
Malignancy | 33.3 (15.2–58.3) | 98.8 (95.6–99.8) | 71.4 (35.9–94.9) | 94.1 (89.5–96.8) |
All data are presented as % (95% CI).
Predictors of Malignant Lesions
A bivariate analysis on the characteristics of benign and malignant lesions is provided in Table 3. There was no significant difference in age at resection based on benign versus malignant lesions (P = .3540). There were no malignant cases associated with a systemic feeding vessel. Bilateral disease and absence of a system feeding vessel were significantly associated with malignancy (P = .0427 and .0203, respectively). Elevated suspicion of a malignancy on preoperative CT was also significantly associated with malignant pathology (P < .0001).
Characteristics of Benign and Malignant Pathology in Postnatally Diagnosed Lung Lesions
Variable . | All Lesions (N = 177) . | Benign (n = 162) . | Malignant (n = 15) . | P . |
---|---|---|---|---|
Male sex, n (%) | 105 (59.3) | 98 (60.5) | 7 (46.7) | .4106 |
Suspected malignancy by CT, n (%) | 7 (4.3) | 2 (1.4) | 5 (33.3) | <.0001* |
Age at preoperative CT, mo, median (IQR) | 7.1 (0.6–58.0) | 5.0 (0.6–58.3) | 19.1 (4.8–53.9) | .1409 |
Age at resection, mo, median (IQR) | 11.1 (3.1–72.0) | 10.9 (2.5–73.9) | 20.7 (7.9–56.1) | .3540 |
Systemic feeding vessel, n (%) | 36 (21.1) | 36 (23.1) | 0 (0.0) | .0427* |
Bilateral disease, n (%) | 3 (1.7) | 1 (0.6) | 2 (13.3) | .0203* |
Clinical symptoms, n (%) | ||||
Respiratory symptoms | 89 (50.6) | 82 (50.9) | 7 (46.7) | .7930 |
History of pneumonia | 54 (30.7) | 50 (31.1) | 4 (26.7) | .9999 |
History of pneumothorax | 6 (3.4) | 4 (2.5) | 2 (13.3) | .0834 |
Asymptomatic | 35 (19.9) | 33 (20.5) | 2 (13.3) | .7385 |
Failure to thrive | 9 (5.1) | 9 (5.6) | 0 (0.0) | .9999 |
Variable . | All Lesions (N = 177) . | Benign (n = 162) . | Malignant (n = 15) . | P . |
---|---|---|---|---|
Male sex, n (%) | 105 (59.3) | 98 (60.5) | 7 (46.7) | .4106 |
Suspected malignancy by CT, n (%) | 7 (4.3) | 2 (1.4) | 5 (33.3) | <.0001* |
Age at preoperative CT, mo, median (IQR) | 7.1 (0.6–58.0) | 5.0 (0.6–58.3) | 19.1 (4.8–53.9) | .1409 |
Age at resection, mo, median (IQR) | 11.1 (3.1–72.0) | 10.9 (2.5–73.9) | 20.7 (7.9–56.1) | .3540 |
Systemic feeding vessel, n (%) | 36 (21.1) | 36 (23.1) | 0 (0.0) | .0427* |
Bilateral disease, n (%) | 3 (1.7) | 1 (0.6) | 2 (13.3) | .0203* |
Clinical symptoms, n (%) | ||||
Respiratory symptoms | 89 (50.6) | 82 (50.9) | 7 (46.7) | .7930 |
History of pneumonia | 54 (30.7) | 50 (31.1) | 4 (26.7) | .9999 |
History of pneumothorax | 6 (3.4) | 4 (2.5) | 2 (13.3) | .0834 |
Asymptomatic | 35 (19.9) | 33 (20.5) | 2 (13.3) | .7385 |
Failure to thrive | 9 (5.1) | 9 (5.6) | 0 (0.0) | .9999 |
P < .05 (Fisher’s exact test).
Multivariable logistic regression was performed by using malignant pathology as the dependent variable and age(s) at first preoperative CT, radiology suspicion of malignancy, bilateral disease, and history of pneumothorax as independent variables (Table 4). Logistic regression confirmed a significant association between suspected malignancy by CT imaging and malignancy as confirmed by pathologic evaluation (odds ratio: 42.15 [95% CI, 7.43–340.4]; P < .0001). Bilateral disease was also a significant predictor of malignant pathology (odds ratio: 42.03 [95% CI, 3.51–995.6]; P = .0041).
Predictors of Malignant Lung Pathology by Multivariable Logistic Regression
Predictor . | Odds Ratio (95% CI) . | P . |
---|---|---|
Age at preoperative CT, mo | 1.00 (0.99–1.01) | .8741 |
Suspected malignancy on preoperative CT | 42.15 (7.43–340.4) | <.0001* |
History of pneumothorax | 7.34 (0.55–62.9) | .0835 |
Bilateral disease | 42.03 (3.51–995.6) | .0041* |
Predictor . | Odds Ratio (95% CI) . | P . |
---|---|---|
Age at preoperative CT, mo | 1.00 (0.99–1.01) | .8741 |
Suspected malignancy on preoperative CT | 42.15 (7.43–340.4) | <.0001* |
History of pneumothorax | 7.34 (0.55–62.9) | .0835 |
Bilateral disease | 42.03 (3.51–995.6) | .0041* |
P < .05 (log-likelihood ratio test).
Discussion
In this report, we used a unique, multi-institutional operative database of 521 pediatric lung lesions to specifically evaluate cases that were not diagnosed with lung lesions in utero. In an era with widespread availability of prenatal ultrasound coupled with ongoing advances in fetal diagnosis and treatment, the management of those who were diagnosed with lung lesions postnatally has largely been ignored in contemporary reviews.30–32 There were several major findings. First, 32% of all pediatric lung lesions are not detected prenatally. Although the missed fetal diagnosis rate may seem abnormally high at first glance, this rate is lower than national data as reported elsewhere.33 Interestingly, the missed fetal diagnosis rate for another major thoracic anomaly, congenital diaphragmatic hernia, was also reported to be 32% on the basis of a registry report.34 Our demographic data, including race and insurance status, do not support the hypothesis that postnatally diagnosed lesions were simply missed fetal lesions that went undetected because of suboptimal prenatal screening. Instead, the relative frequency of the various histopathologic diagnoses in postnatally detected lesions is different. As shown by Fig 1 A–C, CLE is much more common in children who presented with postnatally diagnosed lung masses. Results from our group, among others, have previously revealed that most CLE lesions escape in utero detection by ultrasound, possibly secondary to their isoechoic appearance, absence of mediastinal shift, and atypical fetal growth patterns.35–37 The predominance of CLEs had an impact on symptomatology, in which a greater proportion of children in the postnatally diagnosed group were paradoxically more likely to be symptomatic in early infancy compared with their counterparts with prenatally diagnosed lesions.
A second noteworthy finding from our study was the trimodal age distribution of operative resection based on indications for surgery. Specifically, early surgical intervention within the first 6 months of life was largely because of ongoing respiratory distress. This was followed by a second age cohort of lesions that were identified incidentally and subsequently underwent prophylactic surgery for asymptomatic disease within the first year or two of life. Finally, there were children in a third age cohort, often with a history of pneumonia, who had operative resection after 1 or 2 years of age. The results from this oldest age cohort emphasize the natural history of nonoperative management of many lung malformations (particularly those with intralobar BPS) leading to increased risk of pulmonary infections.13,38
The third and most alarming finding was the incidence of malignant lesions, which was nearly 10% among all postnatally diagnosed masses. The observed incidence of malignancy is slightly higher than data from other published reports.24,39 These differences are likely multifactorial and include differences in sample size, the exclusion of prenatally diagnosed cases in our denominator, and the variability in the pathologic interpretation of specimens.20,24,39,40 On the basis of studies over the past 2 decades, type IV CPAM and cystic PPB have been considered the same pathologic entity.41 Moreover, because there is no pathognomonic molecular marker of PPB, the histopathologic diagnosis relies on other features, such as the presence of immature mesenchymal cells.42 Although case reports of PPB detected in utero have been reported elsewhere,43–45 our data suggest that prenatally diagnosed lesions are at low risk to be cystic PPB and are congruent with work published by others.46,47 On the basis of a recent case-control study, PPB was prenatally diagnosed in only 5% of all confirmed cases.19
As expected, the most common type of lung malignancy in children was cystic PPB, a lesion that was resected at a median age of 10.2 months. These data are similar to work from others,19 including >500 confirmed cases from the International Pleuropulmonary Blastoma Registry (https://www.ppbregistry.org).48 Neonatal cases of cystic PPB have been reported but appear to be uncommon.44 Whether PPB is a de novo cystic tumor or a malignancy that develops within a benign CPAM has not been conclusively demonstrated.22,24,25,40,49,50 However, given the young age and lack of a prenatal diagnosis in our study cohort patients, it seems unlikely that cystic PPB lesions are the result of neoplastic transformation of a preexisting CPAM.51 Children with a malignancy had similar ages at resection when compared with those with CPAMs (median of 20.7 and 17.9 months, respectively). Only a few children with PPB in our series had a positive family history of PPB-associated cancers, which is consistent with the observation made by others that most cases are sporadic.52 Our data also challenge the belief that children with malignancies usually have clinical symptoms because nearly half had asymptomatic disease.19,46 Although cystic PPB lesions have been associated with pneumothoraces, multivariable regression failed to demonstrate previous pneumothorax as a significant predictor of a malignant lesion.
The accuracy of chest CT as a diagnostic tool to help discriminate among cystic and malignant cystic lesions continues to be debated.53,54 We found a positive predictive value of 71% and high odds ratio associated with suspected malignancy by CT, both further supporting the diagnostic utility of this imaging modality when interpreted by an experienced pediatric radiologist. A treatment algorithm, based on clinical criteria and radiologic features such as hyperinflated lung and system feeding vessels, has been proposed to help surgeons decide on whether to resect primary lung lesions.19 However, in 2 recent studies, approximately half of PPB cases were preoperatively misdiagnosed as being CPAMs.46,55 PPB has also been erroneously interpreted as CLE by CT.56 Our data similarly confirmed a relatively low sensitivity for identifying malignant lesions but revealed a negative predictive value of 94%. Factors that may explain the relatively poor sensitivity of CT for malignant lesions likely include a general lack of awareness of cystic PPB as well as overlapping features with macrocystic CPAM.52 Given the desire to further minimize false-negative cases of PPB by CT, we are currently working with pediatric radiologists on a multi-institutional level to estimate interrater reliability in the interpretation of these scans and to further explore what imaging features are most helpful in discerning PPB from benign lung lesions. Finally, multivariable logistic regression analysis revealed that bilateral disease was also a strong predictor of PPB, but this finding was not specific and remains uncommon in children with malignant lung lesions. Based on International Pleuropulmonary Blastoma Registry data, the incidence of bilateral disease and unilateral multifocal lesions in cystic PPB is ∼10%.19,52
The role of preemptive testing for a heterozygous DICER1 germ-line mutation in the evaluation of cystic lung masses has been recommended but is not well studied.57 The high DICER1 mutation positivity rate among our PPB lesions suggests that routine genetic testing in postnatally diagnosed lung lesions may be helpful in identifying suspected cases. However, a recent study has revealed that only 67% of confirmed PPB cases are associated with the DICER1 mutation.18 Our group therefore continues to support a paradigm of aggressive surgical resection, as opposed to clinical observation, in children with lung lesions without documentation of antenatal detection. The low morbidity associated with the operative management of these lesions is demonstrated here and has been reported elsewhere.58,59 Open surgical extirpation in suspected cases may be most prudent to ensure a clear resection margin and to avoid inadvertent aerosolization of tumor throughout the thoracic cavity.57 The use of an impermeable bag for specimen removal after thoracoscopic resection of such lesions has been proposed,60 but the oncologic consequences of minimally invasive approaches remain unknown. Regardless, timely operative management of all cystic PPB lesions is crucial because they are not known to metastasize and are associated with 5-year survival rates of >90% after surgical removal.18,52 In contrast, types II and III PPB, which have a median age of diagnosis of 39 months, are known to metastasize to the brain, bone, and liver, require both surgery and adjuvant therapy, and are associated with a much less favorable prognosis.
Despite the aforementioned findings of this study, several limitations should be acknowledged. First, all data were extracted in a retrospective manner and some variables were limited by missing files and/or data points in the electronic medical record. Second, we did not extract detailed CT data on imaging findings (eg, cyst characteristics, size, distribution) nor did we look for DICER1 genetic testing results in those with benign pathology. Third, despite a strong preference toward resection for nearly all lung malformations among consortium member surgeons, it is possible that some children with pulmonary lesions may not have been captured, especially those who may have never had surgery because of nonoperative management. Finally, our results may not be generalizable to other pediatric referral centers in the United States and elsewhere because of differences in case volumes and case mix.
Conclusions
Among children with primary lung lesions initially detected after birth, PPB appears to be more common than previously thought, occurring in ∼10% of resected lesions. These results strongly caution against routine nonoperative management in this patient population. DICER1 testing may be helpful in identifying suspected cases of PPB given the poor sensitivity of CT for identifying malignant pathology.
Acknowledgments
We acknowledge the data collection efforts of Kevin Johnson and Rodrigo Mon as well as additional integral members of the Midwest Pediatric Surgery Consortium (www.mwpsc.org) for their assistance with this project: Thomas Sato, Daniel von Allmen, Jonathan Kohler, Daniel Ostlie, Jason Fraser, Cynthia Downard, Cheryl Adams, and Sarah Fox.
Dr Kunisaki conceptualized and designed the study, coordinated and supervised data collection, performed data analysis, drafted the initial manuscript, and reviewed and revised the manuscript; Drs Lal, Saito, Fallat, St. Peter, and Fox conceptualized and designed the study, coordinated and supervised data collection, and reviewed and revised the manuscript; Drs Heider, Chan, Boyd, Burns, Deans, Gadepalli, Hirschl, Kabre, Landman, Leys, Mak, Minneci, Wright, and Helmrath conceptualized the study, coordinated data collection, performed data analysis, and critically reviewed the manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: No external funding.
References
Competing Interests
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
Comments