Introduction

Oral submucosal fibrosis (OSF) is an insidious, chronic oral mucosal disease that affects the oral mucosal structure1. OSF begins with a gradual loss of fibroelasticity of the lamina propria, progressing to extensive fibrosis of the oral lamina propria and submucosa, accompanied by epithelial tissue atrophy2. As an important precursor to oral cancer, OSF exhibits a malignant transformation rate of 2.6–7.6% in the Chinese population3,4,5. Additionally, the co-occurrence of OSF and oral potentially malignant disease (OPMD) cannot be ignored. Oral leukoplakia (OLK), which is the most common type of OPMD, is particularly significant in combination with OSF6,7,8,9. OSF with OLK presents a unique and clinically significant scenario where the risk of malignant transformation is synergistically elevated. In contrast, the dual pathology of OSF + OLK exhibits a more complex and insidious nature. While OSF alone disrupts the normal architecture and function of the oral tissues, the addition of OLK introduces epithelial dysplasia with distinct features such as nuclear atypia and architectural disarray, which are hallmarks of precancerous changes. The fibrotic stroma in OSF + OLK not only restricts the normal physiological processes of the tissue but also imposes significant constraints on immune surveillance. This impairment hampers the body’s ability to detect and eliminate abnormal cells, creating a permissive environment for the progression of dysplastic changes.Concurrently, the dysplastic epithelium in OSF + OLK accumulates genetic alterations at an accelerated rate, setting the stage for potential malignant transformation. This accumulation of genetic aberrations, coupled with the fibrotic stroma, disrupts the normal homeostasis of the tissue and fosters a pro-inflammatory microenvironment. This microenvironment is rich in cytokines such as IL-6 and TNF-α, which are promoters of tumor progression and invasion. The reported incidence of OSF combined with OLK (OSF + OLK) ranges from 4.8 to 24.5%. In Taiwan, the malignant transformation rate of OSF + OLK (11.1–18.5%) is significantly higher than that of OSF alone (4.6–7.2%)10,11. This finding emphasizes that the cancer risk is higher when OSF and OLK coexist, providing an important reference for clinical diagnosis and treatment.

The pathological features of OSF mainly involve fibrotic changes in the connective tissue layer; however, the epithelial cell layer is at risk of transforming into malignant tumors, particularly oral squamous cell carcinoma (OSCC)12,13. To establish the transformation pathway from OSF to OSCC, the gold standard generally followed by the academic community involves linking OSF with epithelial dysplasia (an intermediate stage of malignant transformation) through incisional biopsy technology and further exploring its direct association with OSCC through expanded sample analysis14,15. However, incisional biopsy has limitations as a screening tool for evaluating the malignant potential of OSF. 16,17. Thus, many biomarkers have been widely used to predict the malignant transformation of OPMD into squamous cell carcinoma18. Among them, aneuploidy has been supported by solid evidence from follow-up studies, prospective investigations, and clinical trials as a key indicator of malignant transformation19,20. Aneuploidy, which refers to abnormal changes in the number and structure of deoxyribonucleic acid (DNA) or chromosome complements in cells, is a hallmark of malignant cell transformation21,22. Its manifestation at the cellular level––the increase in chromosomal aneuploid cells––often indicates a key early event in cancer development23. To conduct a preliminary diagnostic evaluation of the malignant transformation of squamous epithelial cells, DNA image cytometry (DNA-ICM) is employed as an advanced analytical technology. DNA-ICM can efficiently detect cell morphological changes related to DNA aneuploidy through the fine analysis of Feulgen-stained oral mucosal cell smears, providing a powerful tool for the early detection and intervention of the malignant progression in OPMD24.

This study aimed to explore the potential association between DNA aneuploidy as a biomarker and oral mucosal epithelial dysplasia (OLK) within the context of OSF in a carefully designed, blinded prospective clinical trial. Specifically, this study aimed to verify whether DNA aneuploidy, as a key molecular event, participates in and promotes the progression from OSF to OLK and evaluate whether cases of OSF + OLK carry a higher risk of malignant transformation than OSF alone. To achieve this, we combined the gold standard of histopathological analysis with the advanced detection technology of DNA-ICM to accurately evaluate and quantify the malignant transformation potential of OSF and its accompanying OLK lesions using the dual dimensions of cell morphology and genetics.

Materials and methods

Patients

Patients with OSF who presented to the Oral Medicine Clinic of Second Xiangya Hospital of Central South University between October 2020 and July 2022 were examined for new or review appointments. Stomatologists performed thorough conventional oral examinations. The participants were recruited based on strict inclusion criteria (Table 1). We collected cases of OSF, OSF + OLK, and OSF-concomitant OSCC (OSF + OSCC). Healthy controls with clinically normal oral mucosa were also recruited. All patients received treatment according to the standard management protocols at the Oral Medicine Clinic. To address potential confounding factors, all study participants underwent rigorous screening to exclude individuals with prior treatments that could influence DNA-ICM measurements. Specifically, we confirmed that none of the enrolled patients had received laser therapy, steroid administration, surgical interventions, or other treatments known to affect cellular DNA content prior to sample collection. This exclusion criterion was strictly enforced during the recruitment phase to ensure that the observed DNA-ICM findings could be attributed to the underlying disease processes rather than therapeutic interventions.

Table 1 Inclusion criteria.
Full size table

What’s more, to focus on clinically significant disease stages, we specifically recruited patients with moderate-to-severe OSF as defined by the following criteria:

  1. (1)

    Moderate OSF: Mouth opening 20–35 mm, palpable fibrous bands, epithelial atrophy/moderate hyperkeratosis

  2. (2)

    Severe OSF: Mouth opening < 20 mm, near-complete mucosal fixation, marked epithelial atrophy/severe hyperkeratosis with trismus

These criteria align with stage III–IV classifications in the Khanna-Andrade system, ensuring the study cohort represented advanced disease stages where DNA aneuploidy assessment would have the highest clinical relevance.

Diagnostic criteria

The diagnosis of OSF relies on a comprehensive clinical evaluation and histopathological analysis. Patients with OSF typically have a history of betel nut consumption. Common clinical manifestations include a burning sensation in the oral mucosa, significant dry mouth, recurrent ulcers, and limited mouth opening. These symptoms severely interfere with the patient’s ability to chew, speak, and swallow. At the histopathological level, the progression of OSF begins with a paraepithelial inflammatory reaction, followed by early hyaline degeneration of collagen fibers. In advanced stages, prominent bands of collagen fibers form, with extensive hyalinization extending into the submucosa and a marked decrease in vascular density. Additionally, the epithelium may exhibit pathological changes such as thinning or hyperkeratosis. Coexisting lesions, such as OLK, must be carefully considered during diagnosis. OLK presents with various clinical manifestations, including uniform leukoplakia (characterized by well-defined, flat, and uniform white patches) and heterogeneous leukoplakia (manifested as erythematous areas with nodules or wart-like protrusions). Histopathological features of OLK can range from epithelial atrophy to hyperplasia with or without hyperkeratosis to epithelial dysplasia, and its severity can range from mild to severe. In rare cases, OLK can progress to carcinoma in situ or squamous cell carcinoma, emphasizing the importance of a comprehensive diagnosis.

DNA-ICM

The sample was first placed on a microscope slide, and the cell nucleus was stained dark blue using the Feulgen-thionin staining technique while the cytoplasm remained unstained. The stained samples were fully scanned using an advanced DNA image cytometer from the GEMINI Medical Diagnostic System (Changsha). The DNA content of a cell is expressed as a quantitative index, c, where cells in the G1/G0 phase show a typical 2c (diploid) state, whereas cells in the G2/M phase show 4c (tetraploid) characteristics. Normally, aneuploid cells are absent in the sample, except for cells in the S phase, whose DNA content ranges from 2 to 4c. However, the presence of aneuploid cells is considered a key indicator of carcinogenesis. To ensure accuracy and reliability, a quantitative assessment of DNA content was only performed when at least 300 normalized normal epithelial cell nuclei were present in the sample. We took an approach that did not rely on histological diagnosis and focused on quantifying changes in DNA content of epithelial cell nuclei in the sample. The core of this analysis is to calculate the DNA index (DI), which is the ratio of the DNA content in the nuclei of the tested cells to the DNA content in the standard normal cell nuclei, which is used as the basis for classifying epithelial cell nuclei. Based on the evaluation criteria established in previous literature and the guidelines issued by the European Society for Analytical Cytopathology (ESACP), we set clear thresholds to define different DNA content groups: diploid group (DI between 0.75 and 1.25), Represents DNA content within the normal range; aneuploidy group (DI ≥ 2.3), when more than three nuclei in this group (1% of at least 300 analyzed nuclei) are considered significantly abnormal; Hyperdiploid group (DI is between 1.25 and 1.75), if its proportion reaches 5% of the total number of nuclei, it also indicates abnormality; while in the tetraploidy group (the DI range has overlap in expression, but is usually understood as non- Diploidy and below the aneuploidy threshold, that is, 1.25 ≤ DI < 2.3, the specifics need to be adjusted according to the context to avoid conflict with the definition of aneuploidy), and its evaluation needs to be based on the specific research background and purpose25,26.

To ensure technical reproducibility, rigorous standardization protocols were implemented across all experimental phases. The Feulgen-thionin staining followed ESACP-validated parameters using batch-controlled reagents under fixed temperature (25 ± 1 °C) and duration (45 min). The Leica QWin DNA-ICM system underwent daily calibration with standard HeLa cell pellets (DNA Index 1.8–2.2). Observers completed a 40-h training program with blinded proficiency testing (κ ≥ 0.8 required for certification). Image acquisition parameters were locked (400 × magnification, 25 red/green/blue channels), and automated nuclear segmentation algorithms were validated against manual counting in 10% of samples (R2 = 0.97).

Clinical procedures

Patients were instructed to rinse their mouths with clean water for 1 min to remove any food residue. The lesion area was brushed for 15–20 times using a disposable sterile brush. In patients with OSF, the buccal and tongue mucosae were differentiated. The OSF area was differentiated from the OLK lesion area in patients with OLK + OSF (each patient provided two samples; if either tested positive, the result was considered positive). After the brush biopsy is completed, the brush head is gently twisted off and directly immersed in the cell preservation solution to ensure the integrity of the sample. Subsequently, the sample is prepared according to a rapid process and the necessary staining is performed so that the DNA quantitative analysis can be performed using the high-precision DNA image cytometer of the GEMINI Medical Diagnostic System (Changsha) to ensure the timeliness and accuracy of the results. The results of this quantitative analysis will serve as the basis for preliminary evaluation. Given the authority of histopathological diagnosis, which is regarded as the gold standard for diagnosis, precise surgical biopsies were performed at the brush biopsy site and its adjacent areas under local anesthesia for patients with oral mucosal abnormalities to further verify and refine the diagnosis.

Statistical methods

The required sample size for this study was determined via a prior power analysis. Based on preliminary experimental findings, we assumed a Cohen’s d effect size of 0.6, with a two-tailed significance level (α) of 0.05 and a statistical power (1-β) of 0.8. Calculations revealed a minimum requirement of 28 participants per group. Our actual sample size, with each group comprising 30 or more subjects, exceeded this threshold, thereby ensuring adequate statistical power to detect meaningful effects and supporting the generalizability of the findings.

Result

Diagnostic efficiency of DNA-ICM

This study included 107 histological and exfoliated cell samples, which were divided into four experimental groups: OSF alone group, including 30 samples; OSF + OLK group, including 45 samples; OSF + OSCC group, including 38 samples; and healthy control group, including 30 samples. Among them, the OSF group and OSF + OLK group used samples from previous studies. Figure 1 shows the clinical manifestations of OSF, OSF + OLK and OSCC27. Detailed histopathological analysis and DNA-ICM were performed for each group (Table 2).

Fig. 1
Fig. 1
Full size image

Clinical Manifestations of Patients with Different Oral Conditions. (A) Homogeneous white lesions on the buccal mucosa in a patient with OSF. (B) White plaques on the tongue margin in a patient with OSF + OLK. (C) Abnormal growths on the buccal mucosa in a patient with OSCC.

Table 2 Histopathology and DNA-image cytometry results.
Full size table

Patients without epithelial dysplasia were classified into the negative group. Patients with epithelial dysplasia or cancer were classified into the positive group. The sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) of DNA-ICM for diagnosing epithelial dysplasia and OSCC were as follows:

  • Sensitivity: 88.40% (77.89–94.51%)

  • Specificity: 68.18% (51.28–80.93%)

  • Accuracy: 78.29%

  • PPV: 81.33% (70.33–89.06%)

  • NPV: 78.95% (62.22–89.86%)

Patients with OSF and OSF + OLK without epithelial dysplasia were included in the non-OSCC group. Patients with OSF + OSCC were classified into the OSCC group. The sensitivity, specificity, accuracy, PPV, and NPV of DNA-ICM for diagnosing OSCC were as follows:

  • Sensitivity: 100.00% (88.57–100.00%)

  • Specificity: 68.18% (52.29–80.93%)

  • Accuracy: 84.09%

  • PPV: 73.08% (53.73–84.00%)

  • NPV: 100.00% (85.87–100.00%)

Patients with OSF and OSF + OLK without epithelial dysplasia were included in the non-dysplastic group. Patients with OSF and OSF + OLK with epithelial dysplasia were classified into the dysplasia group. The sensitivity, specificity, accuracy, PPV, and NPV of DNA-ICM for diagnosing epithelial dysplasia were as follows:

  • Sensitivity: 74.19% (55.07–87.46%)

  • Specificity: 68.18% (52.29–80.93%)

  • Accuracy: 71.19%

  • PPV: 62.16% (44.79–77.06%)

  • NPV: 78.95% (62.22–89.86%)

The sensitivity and accuracy of DNA-ICM were the highest for diagnosing severe dysplasia, followed by moderate dysplasia, and the lowest for mild dysplasia (Table 3).

Table 3 DNA image cytometry for diagnosing mild, moderate, and severe epithelial dysplasia.
Full size table

Correlation between epithelial dysplasia and DNA-ICM

Pearson’s chi-square tests were performed to evaluate the correlation between the grading of epithelial dysplasia in OSF, OSF + OLK, and the results of DNA-ICM. Figure 2 shows the examination results of three different types of diseases. The analysis revealed that while the frequency of abnormal DNA content in OSF increased from mild to severe dysplasia, this trend was insignificant (P > 0.05). However, the frequency of abnormal DNA content significantly increased with the severity of dysplasia in patients with OSF + OLK (P < 0.05) (Table 4).

Fig. 2
Fig. 2
Full size image

Results of DNA-ICM analysis at three different ratios. (A-C) An OSF lesion with no obvious abnormal nuclei was diagnosed as non-abnormal proliferation. (D-F) An OSF + OLK lesion with more than 3 nuclei with DI values ≥ 2.3 was diagnosed as moderate abnormal proliferation. (G-I) An OSCC lesion with more than 3 nuclei with DI values ≥ 2.3 was diagnosed as severe abnormal proliferation. (A), (D), (G) are the morphology of aneuploid cells and normal cells (Feulgen staining, lens × 100). (B), (E), (H) are the morphology of aneuploid cells and normal cells (Feulgen staining, lens × 400). (C), (F), (I) are scatter plots showing the number of cells at different DI.

Table 4 Correlation between the epithelial dysplasia grading and DNA content.
Full size table

Differences in DNA-ICM results among disease groups

When comparing the histopathological results among patients with OSF, OSF + OLK, and OSF + OSCC, significant differences in positive results (with epithelial dysplasia or cancer) were observed (Table 5). Patients with OSF + OLK were more likely to exhibit epithelial dysplasia than those with OSF alone, with a risk ratio of 2.315 (95% confidence interval [CI] = 1.138–4.710). Three positive cases are shown in Fig. 3.

Table 5 Analysis of the variability of histopathological results in different diseases.
Full size table
Fig. 3
Fig. 3
Full size image

DNA-ICM analysis results of three different ratios. (A-C) An OSF + OLK lesion with a DI value of more than 3 nuclei ≥ 2.3 was diagnosed as mild abnormal proliferation. (D-F) An OSF + OLK with focal carcinomatosis with a DI value of more than 3 nuclei ≥ 2.3 was diagnosed as severe abnormal proliferation. (G-I) An OSCC lesion with a DI value of more than 3 nuclei ≥ 2.3 was diagnosed as severe abnormal proliferation. (A), (D), (G) are the morphology of aneuploid cells and normal cells (Feulgen staining, lens × 100). (B), (E), (H) are the morphology of aneuploid cells and normal cells (Feulgen staining, lens × 400). (C), (F), (I) are scatter plots showing the number of cells at different DI.

Fisher’s exact chi-square test was subsequently performed to evaluate the relationship between DNA-ICM results across different disease groups. The analysis indicated that the frequency of positive results was the lowest in patients with OSF (33.3%) and the highest in patients with OSF + OSCC (100.0%). The results of DNA-ICM showed significant differences across the groups (P < 0.05), with the positive rate progressively increasing in the following order: OSF, OSF + OLK, and OSF + OSCC. This finding suggests a positive correlation between the rate of aneuploid abnormalities and morphological changes in epithelial cells, progressing from healthy individuals to patients with OSF + OSCC (Table 6).

Table 6 Analysis of the variability of DNA image cytometry results in different diseases.
Full size table

Pearson’s chi-square tests were conducted to analyze the DNA-ICM results among the different disease groups and calculate the relative risk. Significant differences in DNA content were observed between patients and healthy individuals, patients with OSF, those with OSF + OLK, and those with OSF + OSCC (P < 0.05). The relative risk of abnormal DNA content in patients with OSF + OLK compared with those with OSF alone was 1.947 (OR = 1.947, 95% CI = 1.059–3.582). The relative risk of abnormal DNA content in patients with OSF + OSCC compared with those with OSF alone was 4.800 (odds ratio [OR] = 4.800, 95% CI = 2.765–8.332), and compared with those with OSF + OLK, it was 2.407 (OR = 2.407, 95% CI = 1.804–3.212) (Table 7).

Table 7 Differences in DNA image cytometry results between groups.
Full size table

Risk comparison of lesion sites

We collected exfoliated cell specimens from the buccal and tongue mucosae of 308 patients clinically diagnosed with OSF. The positive rate of DNA-ICM for the buccal mucosa was 13.64% (42/308) and 11.36% (35/308) for the tongue mucosa. The Pearson’s chi-square test and relative risk calculations showed no significant differences between these sites (Table 8).

Table 8 Risk comparison of lesion sites between buccal and tongue mucosae of OSF.
Full size table

In patients with OSF + OLK, oral mucosal cells were collected from OLK and OSF lesion areas. The positive rate of DNA-ICM in OLK lesion areas was 60.00% (27/45), whereas that in OSF lesion areas was 17.78% (8/45). Pearson’s chi-square tests and relative risk calculations revealed a significant difference in DNA content between OSF and OLK lesion areas, with the risk of abnormal DNA content in OLK lesions being 2.056 times higher than that in OSF lesions (P < 0.001, 95% CI = 1.402–3.014) (Table 9).

Table 9 Risk comparison of lesion sites between OLK and OSF of OSF + OLK.
Full size table

Additionally, when comparing the DNA-ICM results of OSF lesion areas in patients with OSF + OLK to those in patients with OSF alone, the positivity rate was 17.78% (8/45) in the former group and 21.10% (65/308) in the latter. Pearson’s chi-square test and relative risk calculations showed no significant differences between these groups (Table 10).

Table 10 Risk comparison of OSF lesion sites between patients with OSF + OLK and those with OSF.
Full size table

Discussion

Studies consistently show that abnormal epithelial proliferation increases the risk of malignant transformation in Oral Submucous Fibrosis (OSF). Current clinical assessment primarily relies on histopathological grading of epithelial dysplasia, but accurately identifying biopsy sites reflecting the overall lesion status is challenging due to OSF’s widespread oral mucosal involvement. Severe trismus in patients further complicates biopsy procedures, limiting their feasibility. Historically, OSF was believed to exhibit atrophic epithelial changes, but recent studies reveal dynamic epithelial thickness variations across OSF stages. Sarode’s staging based on mouth opening limitation showed decreasing epithelial thickness with disease progression, with few cases of hyperplasia28. A Hunan study found 37.5% hyperplasia versus 54.1% atrophy, consistent with Pindborg’s report of 71.7% atrophy in 51 OSF cases29. Atrophic epithelium complicates the identification of subtle pathological changes, hampering accurate dysplasia grading and clinical decision-making.

Recent studies indicate a higher-than-expected detection rate of epithelial dysplasia (40.0%) in OSF, improving diagnostic sensitivity30. Early and accurate identification of dysplasia is crucial for targeting high-risk patients for stringent monitoring and intervention, enhancing oral cancer screening and prevention. Patients without dysplasia have a 1.9% malignant transformation rate, highlighting the need for long-term surveillance due to OSF’s widespread lesions and average 37.42-month latency for malignancy31. Quantitative DNA analysis provides an early warning signal, enabling continued surveillance of patients with DNA abnormalities even without evident dysplasia.

Previous studies have revealed that DNA-ICM shows potential in predicting the risk of malignant transformation in OPMD and for early screening of oral cancer, with sensitivity and specificity ranging from 16 to 96.4% and 90–100%, respectively24. However, variations in the definition of high-risk lesions significantly affected the study results. This study further analyzed and discovered that for OSCC, the sensitivity, specificity, and accuracy of DNA-ICM were 100.00%, 68.18%, and 84.09%, respectively, and 74.19%, 68.18%, and 71.19%, respectively, for epithelial dysplasia, with sensitivity increasing with the degree of dysplasia (62.50%, 81.82%, and 100.00%). In summary, the overall sensitivity, specificity, and accuracy of DNA-ICM in detecting oral cancer and epithelial dysplasia were 88.40%, 68.18%, and 78.29%, respectively, demonstrating its clinical value in the early detection and risk assessment of oral cancer.

This study focused on applying DNA-ICM for the early screening and risk prediction of OSF cancer. The study discovered that the coexistence of OSF and OLK is common, accounting for 12.75% (45/353) of the cases. We conducted an in-depth analysis of the correlation between epithelial dysplasia and DNA-ICM results in OSF alone and combination with OLK. The abnormal DNA content in OSF samples increased with the degree of dysplasia, although this difference was insignificant. This finding may be attributed to the limited sample size and absence of cases with moderate dysplasia. In contrast, the frequency of abnormal DNA content increased significantly with the severity of dysplasia in patients with OSF + OLK. Further statistical analysis revealed that from the healthy control group to the OSF, OSF + OLK, and OSF + OSCC groups, the positive detection rate of DNA-ICM showed a gradually increasing trend. Pearson’s chi-square test confirmed that the DNA content in the OSF, OSF + OLK, and OSF + OSCC patient groups was significantly higher than that of healthy individuals, highlighting the high risk of malignant tumors in patients with OSF. Previous population-based longitudinal studies have also demonstrated that patients with OSF had a 29.26 times higher risk of developing oral cancer than those without OSF. Case reports and literature reviews have reinforced the high risk of malignant tumors in patients with OSF + OLK. Compared with patients with OSF alone, those with OSF + OLK showed a higher rate of DNA content abnormalities, but also had a relative risk of malignant tumors, as assessed using histopathology and DNA-ICM. The implementation of ESACP-compliant standardization significantly mitigated technical variability compared to earlier studies. The observed high inter-observer reliability aligns with best practices in diagnostic cytopathology. These measures collectively ensure that observed DNA content differences between OSCC and dysplasia groups reflect biological rather than methodological disparities. Future multi-center validations should adopt comparable standardization frameworks to ensure comparability of results across laboratories. Additionally, studies on OSCC onset in OSF suggest that OSCC has no specific site preference, emphasizing the need for comprehensive clinical monitoring covering the entire oral mucosa.

However, this study had limitations. It is currently in the cross-sectional study stage, making it difficult to evaluate the predictive value of exfoliative cytology for the prognosis of patients with OSF. Future studies will monitor the dynamic changes in DNA-ICM results through regular follow-ups to achieve early detection, diagnosis, and treatment of patients with OSF. Additionally, expanding the sample size, especially to include OSF patients with oral lichen planus, will help explore the impact of combined OPMDs on the malignant transformation of OSF.

Conclusion

This study validates DNA-ICM as a critical tool for assessing OSF malignancy risk and detecting early-stage tumors. By demonstrating high sensitivity and specificity in identifying DNA abnormalities, DNA-ICM emerges as a reliable clinical biomarker for OSF progression. Our findings confirm that OSF patients, particularly those with concurrent OLK, face elevated cancer risks, underscoring the urgent need for intensified surveillance in this high-risk population. To address this clinical gap, we advocate integrating DNA-ICM into routine screening protocols for OSF + OLK patients. This approach enables timely intervention and personalized management strategies, offering a transformative framework to improve oral cancer outcomes. By linking molecular diagnostics to actionable clinical insights, our work provides a novel paradigm for enhancing survival and quality of life in OSF patients.