PPOS is one of the extraskeletal osteosarcomas that originated from the lung with an extremely low incidence and highly invasive potential [3]. Due to its extreme rarity, the clinical characteristics and prognosis of PPOS remain unclear. In this way, we collected the cases with PPOS from the SEER database and conducted a systematic literature review to explore the potential effective treatment and prognostic factors, which helped provide some suggestions for the management of PPOS. The median OS was eight months, and the 1-year OS rate was 30.8%. Besides surgical performance, tumor size and microscopic findings of osteoid were also identified as the independent prognostic factors, and a nomogram was successfully established to predict the prognosis of patients with PPOS with relatively high accuracy. Thus, our findings suggested that surgical resection might be considered for the primary treatment modality. To our knowledge, our research was the first study to demonstrate the clinical characteristics and prognosis of PPOS systematically and put forward a potential prediction model based on a relatively large scale.
PPOS was more common in males (65.3%), higher than the ratio reported before [3, 10]. The median age at diagnosis was 67 years, parallel with other studies [3]. Previously, age was significantly associated with survival [36], while different from our study, we did not find significant correlations between age and prognosis among patients with PPOS. Clinical symptoms might not be notable at an early stage of the disease, and chest pain was the most common as the tumor grew, which was in line with the previous reports [2, 10]. Metastases were only observed in 10 cases in our study, which was correlated with both OS and CSS in the univariable analysis. However, it was not an independent factor in the multivariable analysis. Therefore, the role of metastases in prognosis is required to be further studied because of the limited included cases with the metastasic disease. Our findings might remind us of the typical invasiveness and aggressiveness of PPOS regardless of metastasis.
The tumor size might refer to the local invasiveness, and tumor size > 7 cm indeed led to significantly poor survival, which was also an independent prognostic factor in the multivariable analysis. Previous studies were consistent with our findings that the size might be a critical indicator in determining the prognosis of patients with PPOS [12]. Currently, osteosarcoma was commonly evaluated by pathological grade, while pulmonary osteosarcoma was more recommended to be staged similar to the soft tissue sarcoma of thoracic visceral organs in the American Joint Committee On Cancer staging system, in which primary tumor was described with the depth of invasion and multifocal involvement instead of tumor size [37]. However, our study concluded that tumor size was a significant predictor of survival for patients with PPOS. Therefore, whether tumor size was supposed to be included in the staging system for PPOS required further investigation. Dyspnea was also a common clinical symptom due to the progression of disease and growth of tumor size [10, 18], and it was indicated that the median interval from the onset of clinical symptoms to diagnosis was three to four months, which also seemed to support for the invasiveness [23]. In this way, the growth rate might be a potential prognostic factor [25, 26]; however, it was not documented in detail and unavailable to be investigated properly.
The microscopic findings of the tumor were associated with the prognosis for PPOS. Those without osteoid were associated with a poor OS (HR = 2.11; P = 0.048). Notably, previous research found that there seemed to be some correlation between survival and different histological subtypes in patients with extraskeletal osteosarcoma [7]. In our study, the univariable analysis revealed that calcification was significantly correlated with survival. Generally speaking, calcification found in a mass on CT images was considered a benign sign, even though calcified findings were considered the specified characteristic of PPOS [2, 12]. Therefore, microscopic findings and radiographic manifestations might be significantly associated with the prognosis of PPOS, which required further validation. Furthermore, the microscopic findings could also be available for those who had not received surgical resection, as pathological examination could be conducted via transbronchial lung biopsy or needle biopsy. In addition, the presence of pleural effusion was associated with survival in univariable analysis but not identified as an independent prognostic factor in multivariable analysis. Due to the lack of details on whether these effusions were malignant, this variable was not included for further analysis.
Due to its rarity, only surgery was recommended as a potentially effective treatment for PPOS, and the role of chemotherapy and radiotherapy remains controversial even though chemoradiotherapy seemed to decrease the risk of distant metastasis and prolong the survival [2, 6, 7, 12, 13]. In our study, the Cox proportional hazard model showed that patients with surgery would have better OS (HR = 0.20, P < 0.001), which was in line with other studies [1, 2, 36]. The prognosis of patients with PPOS was extremely unsatisfactory, while those who received surgical resection might achieve nearly twice OS than those without surgery (ten versus five months). Thus, even though the potential invasiveness and aggressiveness were noted, surgery was still recommended. Furthermore, extensive radical resection was more recommended to remove the primary tumors and achieve a better local control [2, 7, 12]. Chapman reported a 33-year-old woman without a smoking history, who received surgery and adjuvant chemotherapy, survived 42 months postoperatively [29]. However, compared with en bloc resection or extensive surgery, inadequate resection was referred to as poor survival [22]. Therefore, it seemed essential to achieve a radical resection as possible when performing the first attempt of surgical resection. Radiotherapy was not effective, as shown in our findings, which was consistent with the previous studies [21, 30], while radiotherapy might be helpful in patients with distant metastasis [13, 33]. Moreover, even though the potential benefits of chemotherapy had been demonstrated in 266 patients with extraskeletal osteosarcoma [38], chemotherapy might not lead to a better OS in this study (P = 0.276).
What’s more, it is profitable to construct a prediction model to predict the prognosis of patients with rare diseases so as to help the clinicians make the right decision, and in this way, we constructed a nomogram, which could effectively assess the survival of those with PPOS (C-index, 0.771; AUC, 0.818). Unlike previous studies focusing on extraskeletal osteosarcoma [36], we merely studied the malignancy originating from the lung. To our knowledge, this is the first prediction model for PPOS, via which we might effectively predict the prognosis of patients with PPOS.
Our study had several limitations. As the data were collected from the reported cases and SEER database, the details of comorbidities, treatment contents and completeness of surgical resection were unavailable, which might lack persuasiveness to some extent. Secondly, due to the low incidence of PPOS, the included sample scale was not satisfactory, and there was no external cohort to further validate our conclusions. Moreover, including cases from the SEER database and case reports might create some data heterogeneity and introduce numerous biases. Therefore, our conclusions should be interpreted with caution and further multicenter studies are badly needed to better elucidate the clinical characteristics and prognosis of PPOS.