"How much time do I have, doctor?" This is an important question for cancer patients, and one that clinicians often struggle to address. For patients and clinicians, meaningful prognostic information is essential for making informed treatment decisions. In the present study, we developed and validated a simple and reliable tool to predict the survival 1-,2-, and 3-year in lung cancer patients under immunotherapy. We identified nine independent prognostic factors (age, TNM stages, surgery, radiation, KPS, histology, multidrug.therapy, D-dimer and ALB) that were further included in the nomogram based on their statistical significance. According to the total points in nomogram, the cutoff values were determined to stratify patients into different risk subgroups, which can be more useful in assisting patient counseling and guiding treatment decision. Additionally, based on the nomogram model, we have developed a user-friendly online calculator, enabling clinicians to easily use.
Prior studies have indeed utilized nomograms to predict patient outcomes in the context of lung cancer immunotherapy.(He et al. 2022) However, Our study is unique in that it focuses specifically on a combination of factors such as age, TNM staging, surgery, radiotherapy, KPS, histology, multidrug therapy, D-dimer, and ALB levels, and we included all ICIs marketed both domestically and internationally during the study period. Previous studies such as the one by He B.X. and colleagues, have largely focused on imaging-based scores using advanced techniques like deep learning(He et al. 2022). While these are valuable, they often lack consideration of certain clinical or biological factors that our study addresses. Other studies may be limited by focusing only on specific types of ICIs, such as PD-1 inhibitors, or only including foreign drugs without considering those available and commonly used in specific regions like China. Thus, while nomograms have been used in the context of lung cancer immunotherapy, our approach adds a more comprehensive and regionally relevant model to this existing body of work(Yuan et al. 2021; Wang et al. 2022; Li et al. 2022).
We included people ageing from 29 to 84 years who were treated with ICIs. Although age was not statistically significant in univariate analysis, we still included age in the prediction model, considering that age was associated with survival in cancer patient(Bates et al. 2022) .
Our results suggest that stage Ⅳ tumor is a risk factor for lung cancer OS, which is easy to understand, and that OS is certainly shorter in stage Ⅳ patients than in stage Ⅰ-Ⅲ patients. This result is consistent with previous research results(Ouyang et al. 2022).
Current primary lung cancer treatments encompass surgery, radiotherapy, and pharmacotherapy(Duma, Santana-Davila, and Molina 2019). In addition to drug therapy, surgery can improve the survival of lung cancer, and radiotherapy can also bring significant benefits to patients(Han and Lee 1988). In addition, the baseline and long-term survival of patients who had access to surgery and radiotherapy were better than those who only received medical treatment. This is consistent with the fact that surgery and radiotherapy were protective factors for total production in this study.
The KPS assesses patients' functional status. A KPS of more than 80 indicates that patients can take care of themselves(She et al. 2019). When patients' KPS is below 70, ICIs should be used with caution(Thompson et al. 2022).Our study corroborates the idea that KPS is a protective factor for OS, with higher KPS scores implying longer patient survival.
Although there is no statistical difference in histology, the survival of different types of lung cancer receiving immunotherapy is different. At present, the efficacy of immunotherapy for NSCLC is better than that for small cell lung cancer (Mamdani et al. 2022). Therefore, we also included histology in the model.
Now, chemotherapy combined with immunotherapy is the standard treatment for lung cancer, except for patients with high PD-L1 expression (Maung et al. 2020). Our findings indicate that combining chemotherapy with immunotherapy, compared to immunotherapy alone, enhances progression-free survival and OS for patients(Wang et al. 2021).
D-dimer, a product of fibrin degradation, reflects fibrinolysis function. Elevated D-dimer levels signify hypercoagulability and secondary hyperfibrinolysis, with clinical detection primarily used in diagnosing venous thrombus embolism, deep venous thrombosis, and pulmonary embolism(Lei et al. 2022). Recent data reveals that cancer patients exhibit a 4–7 times higher thrombosis risk compared to non-cancer patients(Fernandes et al. 2019). Among the death factors of tumor patients, thrombosis is the second most important factor after tumor recurrence. Therefore, elevated D-dimer is a risk factor for survival.
Tumor is generally a wasting disease, and the nutritional level of tumor patients is closely related to their survival, ALB is a kind of nutrient in the body(Polański et al. 2021). Usually, the nutritional level of end-stage patients is very poor, and some patients need to constantly supplement ALB. Therefore, ALB is a protective factor for long-term survival.
In addition, we noted that immune-related indicators, such as lymphocyte counts and CD4/CD8 ratio, did not show statistical significance in our univariate analysis. A number of potential reasons could account for this finding. It might be attributed to factors such as the size of our sample, presence of comorbidities, treatment modalities, or influence of other biomarkers. However, it is important to note that this does not negate the relevance of these immune-related indicators in predicting responses to ICIs therapy. Especially considering the mechanism of action of immune checkpoint inhibitors which involve T lymphocyte function, the role of lymphocyte subsets cannot be undermined. Future research should further investigate these immune-related indicators and attempt to understand why they didn't show statistical significance in our model.
Calibration and ROC curves displayed strong predictive performance, while the DCA curve indicated the nomogram's clinical utility. There are several limitations that need to be addressed. First, we lacked valuable data (eg. gene mutation, PD-L1 expression status), which may be an important predictor for immunotherapy. The lack of such significant information might impact the accuracy of our predictive model However, as retrospective data, it is not directly accessible in our electronic medical records and not all patients will undergo genetic testing or PD-L1 testing. However, the variables we include now are routinely tested by patients and are relatively easy to obtain. And these factors still have predictive value. Second, our study included single-center data collection, a relatively small sample size, and the absence of external validation. In subsequent studies, we will encourage the performance of multicenter studies to further verify the reliability and applicability of the nomogram. Third, since the CTLA-4 inhibitors ipilimumab was approved for lung cancer treatment by China Food and Drug Association until October 2021, and has low accessibility because of its high economic burden in China. Our study did not include patients treated with CTLA-4 inhibitors, possibly introducing bias. Finally, although the advantage of the nomogram is its convenience in clinical practice, it will be considered other potential prognostic variable, such as ethnic, occupation, smoking history and comorbidity, which may enhance the predictive performance.