Tyrosine Kinase Inhibitor Therapy in Pediatric Sarcoma: Prognostic Implications of Pulmonary Metastatic Cavitation

Abstract

Purposes:

This study aims to ascertain the prevalence of cavitations in pulmonary metastases among pediatric and young adult patients with sarcoma undergoing tyrosine kinase inhibitor (TKI) therapy and assess whether cavitation can predict clinical response and survival outcomes.

Methods:

In a single-center retrospective analysis, we examined chest CTs of 17 patients (median age 16 years; age range 4-25 years) with histopathologically confirmed bone (n=10) or soft tissue (n=7) sarcoma who underwent TKI treatment for lung metastases. The interval between TKI initiation and the onset of lung nodule cavitation and tumor regrowth were assessed. The combination of all imaging studies and clinical data served as the reference standard for clinical responses. Progression-free survival (PFS) was compared between patients with cavitating and solid nodules using Kaplan-Meier survival analysis and log-rank test.

Results:

Five out of 17 patients (29%) exhibited cavitation of pulmonary nodules during TKI therapy. The median time from TKI initiation to the first observed cavitation was 79 days (range 46-261 days). At the time of cavitation, all patients demonstrated stable disease. When the cavities began to fill with solid tumor, 60% (3/5) of patients exhibited progression in other pulmonary nodules. The median PFS for patients with cavitated pulmonary nodules after TKI treatment (6.7 months) was significantly longer compared to patients without cavitated nodules (3.8 months; log-rank p-value = 0.03).

Conclusions.

Cavitation of metastatic pulmonary nodules in sarcoma patients undergoing TKI treatment is indicative of non-progressive disease and significantly correlates with PFS.

Introduction

Metastatic sarcomas have a poor prognosis, with bone sarcomas exhibiting 5-year survival rates ranging from 33% to 45%, and soft tissue sarcomas exhibiting rates of 19% to 47% in children and young adults.¹ The lung is the most common site for metastases at the time of initial diagnosis and at the time of disease recurrence.² Over 90% of relapsed bone sarcomas and up to 60% of relapsed soft tissue sarcomas are associated with metastases in the lung.^3-7 Systemic therapy is generally used for metastatic disease recurrence except in carefully selected patients with specific sarcoma subtypes and clinical features that suggest a surgery-only approach. Over the last decade, several clinical studies have shown that antiangiogenic tyrosine kinase inhibitors (TKIs) can delay progression of metastatic sarcoma following standard multimodal therapy.^8-13 TKIs can lead to distinct changes of normal and neoplastic tissues on medical imaging studies. Several authors described that TKIs can cause a spontaneous pneumothorax in up to 13% of patients in clinical trials^8-10,12,13 and up to 25% of real world studies.^14-16 It was noted that cavitation of pulmonary nodules may be associated with the development of a pneumothorax.^15,17 However, the frequency of tumor cavitation and its association with clinical response has received surprisingly limited attention thus far.

To close this gap, our study aimed to determine the frequency of cavitating metastatic pulmonary nodules in children and young adults with sarcoma who received TKIs, and to evaluate whether cavitation can predict clinical response and survival outcome.

Materials and Methods

Patients

This single-center retrospective study was approved by the institutional review board. The analysis of de-identified data was permitted and informed consent was waived. The inclusion criteria were 1) histologically proven newly-diagnosed or recurrent sarcoma, 2) received at least one cycle (30 days) of TKI, and 3) age younger than 25 years old when TKI was given. We identified 24 children and young adults who met the inclusion criteria between January 2008 to February 2023. Patients were excluded if they had no lung metastases (n=3), if CT imaging studies before and during/after TKI therapy were not available (n=3), or if date of initiation and discontinuation of TKI was not available (n=1).

Radiologic evaluation

Two board-certified radiologists (W.M. and L.C.A.) with 5 and 7 years of experience, respectively, reviewed 71 chest CT scans of 17 patients in consensus. These included 17 CT scans at baseline, before start of TKI therapy, and 54 scans in the same patients during and after TKI therapy. At the time of interpretation, the readers were aware that the patients had pulmonary metastases and were receiving TKIs, but they were blinded to other clinical data. CT scans of the chest (without [n=62] or with [n=9] intravenous contrast agent) were reconstructed at 2-mm slice thickness to evaluate the pulmonary nodules. For all scans, the reviewers recorded the number and size of pulmonary nodules with a diameter ≥ 3 mm in longest diameter, the presence/absence of cavitation, and the presence/absence of a pneumothorax. A cavitation was defined as a gas-filled space seen as a lucency or low attenuation area within at least 1 pulmonary nodule.¹⁸ The cavitated nodule can have either a thick, irregular wall, or a cyst-liked thin wall.

Standard of reference

The clinical response was assessed based on all available clinical imaging data and included CT (n=71), whole-body 18F-fluorodeoxyglucose (FDG) PET/MRI (n=30), and MRI (n=19). The disease response was assessed based on PERCIST 1.0 if there were lesions with increased metabolic activity on 18F-FDG PET/MRI. Otherwise, the response assessment was based on RECIST 1.1. We divided patients into two groups: progressive disease (PD) and non-progressive disease (non-PD; partial response or stable disease). Disease progression was defined as at least a 30% increase in peak standardized uptake value corrected for lean body mass (SULpeak) on 18F-FDG PET/MRI according to PERCIST 1.0, at least a 20% increase in the sum of the disease measurements for measurable lesions on CT and MRI according to RECIST 1.1, or the appearance of new lesions.¹⁹ The median time between pre-treatment scans to the first dose of TKI was 15 days (range, 0 – 60 days). The progression-free survival (PFS) was defined as the interval between the date of initiation of TKI until progression of disease while on TKI as the end date (n=15). For patients who stopped TKI therapy due to patient decision (n=1) and adverse event (n=1), the date of disease progression prior to initiation of new anti-cancer therapy was the end date of PFS.

Statistical analysis

Descriptive statistics were used to summarize the patients’ demographic characteristics, time to developing cavitation, size of each cavitary pulmonary metastasis, presence/absence of pneumothorax and emergence of new solid components within the cavity. Kaplan-Meier plots were used to illustrate the difference in survival times between patients whose nodules did or did not demonstrate cavitation. The log-rank test was used to test equality of survival functions between patients who did and did not develop pulmonary nodule cavitation. Statistical analyses were performed by an expert statistician (T.L.) using STATA software (version 17.0; StataCorp, College Station, Tex), assuming significant differences for p-value < 0.05.

Results

Patient characteristics

Seventeen patients (10 bone sarcoma, 7 soft tissue sarcoma) with pulmonary metastases were included in the study: These included 12 males and 5 females with a median age of 16 years (range, 4 to 25 years) at initiation of TKI therapy. Fourteen patients (82%) were recurrent diseases of which primary tumors were surgically resected and 3 patients (18%) demonstrated persistent primary tumors with disseminated diseases. Eleven patients (65%) had metastatic diseases in the lungs and other organs (bone, pleura, muscle, lymph node, kidney, liver, pancreas) and 6 patients (35%) had only pulmonary metastases. The median duration of TKI treatment was 5.4 months (range, 1.6 to 11 months). Further patient, tumor, and treatment data are provided in Table 1.

TABLE 1.

Patient characteristics and demographics

Characteristic	Value
Age (year)
Median (interquartile range)	16 (13 – 19)
Range	4 – 25
Sex
Male	12 (71%)
Female	5 (29%)
Ethnicity
Non-Hispanic origin	9 (53%)
Hispanic origin	8 (47%)
Race
White	6 (35%)
Other	11 (65%)
Bone sarcoma	10 (59%)
Osteosarcoma	9 (53%)
Ewing sarcoma	1 (6%)
Soft tissue sarcoma	7 (41%)
Rhabdomyosarcoma	3 (18%)
Non-rhabdomyosarcoma (alveolar soft part sarcoma, myofibrosarcoma, synovial sarcoma, undifferentiated pleomorphic sarcoma)	4 (23%)
Status of disease
Recurrence	14 (82%)
Primary diagnosis	3 (18%)
Number of pulmonary nodules/masses
Median	3
Range	1 – 131
Extrapulmonary disease
Yes	11 (65%)
No	6 (35%)
Tyrosine Kinase Inhibitor
Pazopanib	5 (29%)
Regorafenib	3 (18%)
Cabozantinib	3 (18%)
Sorafenib	2 (11%)
Sorafenib and everolimus	3 (18%)
Anlotinib	1 (6%)
Duration of treatment (month)
Median (interquartile range)	5.4 (2.6 – 6)
Range	1.6 – 11

Unless otherwise indicated, values are n (%)

Pulmonary nodule cavitation and treatment response

All patients demonstrated solid metastatic pulmonary nodules on their baseline chest CT. After TKI administration, 12 patients had period of non-PD and 5 patients experienced PD at the first follow up examination (Fig 1).

Figure 1:

Flowchart demonstrates the characteristics of pulmonary metastases and clinical course of pediatric patients with bone or soft tissue sarcoma who underwent tyrosine kinase inhibitor (TKI) therapy. PD, progressive disease.

Five of the 12 patients with non-PD (29% of the entire cohort) demonstrated cavitation of at least one pulmonary nodule after TKI therapy (4 osteosarcoma and 1 Ewing sarcoma; Fig 2). The number of nodules that cavitated ranged from 1 to 3 nodules per patient and the longest diameter of nodules ranged from 0.3 to 2.6 cm. The median time between start of TKI therapy and first observed cavitation was 79 days (range, 46 – 261 days). All patients who demonstrated cavitation of at least one pulmonary nodule demonstrated stable disease (n=3) or partial response (n=2). Follow up imaging studies demonstrated increasing solid components within the cavities in 3 of 5 patients. All of these patients demonstrated an increased size of pulmonary nodules on the same imaging study (Fig 3), or a subsequent imaging study (Fig 4). The cavities remained unchanged in 2 patients despite worsening of other metastatic sites. Supplemental Table S1 summarizes the characteristics and cavitary lesions of the 5 patients.

Figure 2:

An 18-year-old male with metastatic Ewing sarcoma. Axial chest CT at baseline (A) shows a large pulmonary nodule in the left upper lobe (large arrow) and multiple smaller pulmonary nodules in the right lung (small arrows). Maximum intensity projection of 18F-fluorodeoxyglucose (FDG) PET demonstrates high metabolic activity of the nodules in the bilateral lungs and multiple bone metastases on the axial and appendicular skeleton (B). After 2 months of the treatment with cabozantinib, chest CT (C) shows developing cavity of the nodule in the left upper lobe (large arrow) and decrease in size of the nodule in the right lung (small arrow). All pulmonary nodules and bone metastases demonstrated decreased FDG uptake on 18F-FDG PET (D), indicating response to the therapy.

Figure 3:

A 12-year-old male with recurrent osteosarcoma. Axial chest CT at baseline (A) shows a pulmonary nodule in the right middle lobe (arrow). Follow up chest CT at 1.5 month after initiation of regorafenib (B) demonstrates right pneumothorax and decrease in size of the pulmonary nodule (arrow), however, no cavitation was detected. The patient was asymptomatic. On 3.5-month follow up scans (C), the right middle lobe nodule shows cavitation (arrow) and the right pneumothorax resolved spontaneously. After 6 months of therapy (D), the disease progressed and the cavitary nodule is completely filled with soft tissue (arrow). Recurrence of right pneumothorax is noted (arrowheads).

Figure 4:

A 13-year-old male with recurrent osteosarcoma. Axial chest CT at baseline (A) shows a nodule in the left lower lobe (top row; black arrow) and right lower lobe (bottom row; white arrow). At 1.5 month after treatment with sorafenib and everolimus (B), the nodule in the left lung is completely cavitated (arrow) and the right lower lobe nodule is stable in size (arrowhead). At 3-month follow up (C), the cavitating nodule in the left lower lobe demonstrated internal tumor regrowth (black arrow) and the right lower lobe nodule is not significantly changed (white arrow). The patient was considered to have stable disease at the time. On 5.5-month follow up scans (D), the significantly enlarged right lower lobe nodule (white arrow) and a new nodule (not shown) confirmed progression of disease. The patient underwent wedge resection of the left lower lobe nodule and metastatic high-grade osteosarcoma was confirmed by histopathology.

The remaining 12 patients (71%) who did not show cavity formation comprised 7 patients (4 osteosarcoma, 3 soft tissue sarcoma) with non-PD for 3.4 to 9.9 months, and 5 patients (1 osteosarcoma, 4 soft tissue sarcoma) with progression of disease on the first follow up imaging study (ranging from 1.8 to 3.6 months) due to new metastatic lesions (n=4) or significant increase in size of the pre-existing metastatic lesions (n=1).

Pneumothorax

We observed a pneumothorax in 3 of the 5 patients with pulmonary nodule cavitation (60%) but none of the patients without cavitation. Two of the 3 patients (patient 4 and 5; Supplemental Table 1) with pneumothorax had cavitary changes of subpleural nodules that preceded the pneumothorax. One of 2 patients with cavitating nodules that were in central location also had pneumothorax. The pneumothorax occurred prior to detected cavitation in one patient (Fig 3; patient 3; Supplemental Table 1). No imaging evidence of pleural metastasis was observed in patients with pneumothorax.

Survival analysis

Overall, 4-month PFS of the patients after the initiation of TKI was 59% and median PFS was 5.5 months (95% confidence interval [CI] 3.9 – 6.7). The 4-month PFS and median PFS were 100% and 6.7 months (95% CI 4.6 – 10.8) for patients with cavitating nodules, and 42% and 3.8 months (95% CI 2.9 – 5.7) for patients with non-cavitating, solid nodules (log-rank p-value 0.03). The Kaplan-Meier curve is shown in Figure 5.

Figure 5:

Kaplan–Meier curves for progression-free survival of patients with and without developing cavitating pulmonary nodules under tyrosine kinase inhibitor therapy (p-value = 0.03).

Discussion

We demonstrated that cavitating pulmonary nodules are common in children and young adults with metastatic sarcoma who received TKIs. All patients with cavitating nodules had disease that was, at least initially, stable or shrinking in response to therapy as long as the cavitation was observed.

In our study, pulmonary nodule cavitation occurred in 29% of sarcoma patients. Our result agrees with the prior largest study which reported tumor cavitation in 27% of pediatric patients with various tumor types, including sarcomas, Wilms tumor, and renal cell carcinoma, treated with sorafenib, bevacizumab, and low-dose cyclophosphamide.¹⁵ We found that cavitation occurred in a few pulmonary nodules of patients similar to previous case reports of cavitary pulmonary metastases in osteosarcoma patients.^20,21 Our finding differs from previous report of wide-spread cavity development in Wilms tumors with pulmonary metastases following antiangiogenic therapy.¹⁵ We did not observe tumor cavitation in soft tissue sarcoma, although tumor cavitation has been reported in 30% of adults with soft tissue sarcoma treated with pazopanib.¹⁶ The majority of soft tissue sarcomas that showed cavitation in adults was undifferentiated pleomorphic sarcoma which is rare in children.^16,22 The most common soft tissue sarcoma in pediatric population, and also in our study, is rhabdomyosarcoma which has demonstrated poor response to pazopanib.²³ In addition, we found that most of the patients with early progression had soft tissue sarcoma, supporting evidence in the literature that soft tissue sarcomas may less sensitive to TKIs and therefore less likely to develop cavitation than bone sarcoma.

Most of TKIs target vascular endothelial growth factor receptors (VEGFR) that promote angiogenesis (Supplemental Table S2). Inhibition of vascular endothelial growth factor induced decreased vascularization and tumor necrosis as shown by preclinical studies.^24,25 Alteration of tumor blood volume after treatment with TKI is also proven by clinical studies.^26,27 In our study, we found that all patients showed lack of tumor progression at the time cavities occurred with one patient demonstrating metabolic response. Thus, cavitation may indicate favorable therapeutic response even though the lesion may not fulfill criteria for RECIST-defined target lesions. In patients receiving VEGFR inhibitor and epidermal growth factor receptor (EGFR) TKI therapy for primary lung cancer, an alternative measuring method has been developed to assess cavitating tumor by measuring only the solid tumor components, and excluding the air component of the cavity.^28,29 In addition, it is important to note that central filling of a cavitating nodule with solid tumor may indicate tumor progression even though the overall diameter remains the same.²⁸ We found solid filling in 60% of the patients with cavitary lesions when the disease progressed. We suggest that morphologic changes such as nodule cavitation and later filling of cavitating nodules be further studied to evaluate their potential contributions to response assessment beyond the nodule size changes used in current guidelines.

In the past, spontaneous pneumothorax has been reported up to 11% of patients with sarcoma, either before or after treatment such as radiation therapy and cytotoxic chemotherapy.^20,30-34 After introduction of antiangiogenic therapy including TKIs, an increased number to 25% of cases with spontaneous pneumothorax have been noted during cancer therapy.^{10,14-16,35,36} Prior studies suggested an association between pneumothorax and cavitation of pulmonary metastasis.^15-17,35 Analyses of patients with tumor cavitation reported a pneumothorax in 75% of pediatric patients with solid tumors, 93% of patients with osteosarcoma, and 83% of adults with soft tissue sarcoma.^15,16,35 Our observation of a pneumothorax in 3 of 5 patients with cavitating nodules in children and young adults with sarcoma is consistent with past reports. The incidence of pneumothorax is slightly higher in the patients who had subpleural cavitating nodules than the patients with cavitating nodules in central location (67% versus 50%) in our study. However, no significant association between the subpleural location of cavitating nodules and pneumothorax has been found in the previous report.¹⁵ Pleural metastasis did not seem to relate with pneumothorax since none of patients with pneumothorax had pleural metastasis in current study. Development of pneumothorax is an important adverse event and can occur prior to or following the development of cavitation. Thus, the clinical care team should be aware of this possible adverse event after TKI

There are several limitations to the present study. Firstly, the investigation was confined to patients receiving TKI therapy at a single center, which consequently restricted the sample size. Given the rarity of pediatric cancers and the emerging role of TKIs as a therapeutic modality, most clinical trials have focused on adult patients. Consequently, our study serves as a preliminary exploration to highlight the disparity in PFS between patients with and without tumor cavitation. The survival outcomes observed in our cohort align with the median PFS reported in the literature, ranging between 3.6 and 6.7 months.^8-13 Despite the small sample size, our findings suggest that patients presenting with tumor cavitation exhibited significantly longer PFS compared to those without cavity development. This association between cavitation and extended survival necessitates validation with multifactorial analysis through larger prospective patient cohorts, likely mandating a multi-center collaborative effort. Additionally, our study underscores the implications of tumor regrowth within a cavity that had previously developed in a pulmonary nodule, which was strongly associated with disease progression in the near term. If further study confirms this association, filling in of a pulmonary nodule cavity may be properly regarded as evidence of tumor progression, even in the absence of growth of the nodule.

In conclusion, cavitation of metastatic pulmonary nodules is common in sarcoma patients who receive TKIs, carries a substantial risk of pneumothorax, and may be an indicator of therapeutic efficacy even in the absence of nodule shrinkage. Future research endeavors should investigate whether tumor type influences the likelihood of cavitation and how reliably cavitation and subsequent filling of these cavities correlate with therapeutic response.

Abbreviations:

CT

Computed tomography

CI

Confidence interval

EFGR

Epidermal growth factor receptor

FDG

Fluorodeoxyglucose

MRI

Magnetic resonance imaging

PD

Progressive disease

PERCIST

PET Response Criteria in Solid Tumors

PET

Positron emission tomography

PFS

Progression-free survival

RECIST

Response Evaluation Criteria in Solid Tumors

SULpeak

Peak standardized uptake value corrected for lean body mass

TKI

Tyrosine kinase inhibitor

VEGFR

Vascular endothelial growth factor receptors