Skip to main content

Establishment and validation of nomogram for predicting immuno checkpoint inhibitor related pneumonia



Cancer is one of the main causes of death worldwide. Although immunotherapy brings hope for cancer treatment, it is also accompanied by immune checkpoint inhibitor-related adverse events (irAEs). Immune checkpoint inhibitor pneumonia (CIP) is a potentially fatal adverse event, but there is still a lack of effective markers and prediction models to identify patients at increased risk of CIP.


A total of 369 cancer patients treated between 2017 and 2022 with immune checkpoint inhibitors at Shengjing Hospital of China Medical University and Liaoning People's Hospital were recruited for this study. Independent variables were selected by differences and binary logistic regression analysis, and a risk assessment nomogram was constructed for CIP risk. The accuracy and discriminative abilities of the nomogram were evaluated by calibration plots, receiver operating characteristic curves (ROCs) and decision curve analyses (DCAs).


Binary logistic regression analysis showed that smoking history, acute phase proteins [interleukin (IL-6) and C-reactive protein (CRP)], CD8 + T lymphocyte count and serum alveolar protein [surface protein-A (SP-A) and Krebs Von den Lungen-6 (KL-6)] were significantly associated with CIP risk. A nomogram consisting of these variables was established and validated by different analyses.


We developed an effective risk nomogram for CIP prediction in immune-checkpoint inhibitor administrated cancer patients, which will further assist early detection of immunotherapy-related adverse events.

Peer Review reports


Cancer is a serious threat to human health and life and is the second most important cause of death after heart disease. In 2022, nearly five million new cancer cases and three million cancer deaths are projected to occur in China [1]. Immune checkpoint inhibitors (ICIs) have changed the pattern of tumour treatment; however, they are accompanied by multisystem ICI-related adverse events (irAEs). irAEs occur in up to 70% of patients administered Programmed Death-(Ligand) 1 inhibitors [2]. Most irAEs, including cutaneous, liver, gastrointestinal, and endocrine adverse events, are controllable. However, some irAEs occurring in the heart, lungs, liver, and nervous system are potentially fatal and life-threatening.

Immune checkpoint inhibitor pneumonia (CIP) is a kind of irAE that occurs in the lungs, and the main lesion is interstitial pneumonia originating in the lower lobes of both lungs. CIP has been reported to have a morbidity rate of 3–5% and a mortality rate of 10–17% and it may be higher in patients with lung cancers [3]. Its main symptoms are often nonspecific, including dyspnoea, cough, fever and chest pain. CIP occurs mainly in the first 6 months after starting treatment [4]. Glucocorticoids can be used for CIP treatment, and are usually effective, but approximately 1/4 of patients eventually relapse [5]. As a serious irAE, CIP has a hidden onset. It is PD-1/PD-L1 dose-independent and lacks effective early predictive and models. After therapy, about one-fourth of CIP patients will develop recurrence [6].

Pulmonary surfactant is an important lipoprotein complex that is mainly produced by alveolar epithelial cells (AECs). Pulmonary surfactant plays an important role in alveolar air exchange and participates in pulmonary immune regulation and antiviral infection. Surfactant protein A (SP-A) and surfactant protein D (SP-D) are large, soluble, hydrophilic proteins that are important components of the pulmonary surfactant. The production of pulmonary surfactant may increase after alveolar epithelial cell injury. However, the markers of alveolar epithelial cell lesions lack specificity and are changed in many diseases, such as systemic sclerosis-associated interstitial lung disease [7] and idiopathic pulmonary interstitial fibrosis [8]. For cancer, Hasegawa et al. reported that SP-D inhibited the proliferation and motility of NSCLC by binding to epidermal growth factor receptor (EGFR) directly [9], so serum SP-D levels may predict the response to EGFR-tyrosine kinase inhibitors (EGFR-TKIs) [10].

Krebs von den Lungen-6 (KL-6), which is classified as one of the human Mucin 1 antigens, is a high molecular weight glycoprotein expressed predominantly by AECs, bronchial glandular epithelial cells and bronchial gland cells [11]. KL-6 was initially studied as a cancer biomarker. When the expression of KL-6 in the serum and tumour tissue of patients with non-small cell lung cancer is increased, they have a poor prognosis [12, 13] and worse curative effect of EGFR-TKIs [14]. KL-6 is only expressed at low levels in normal lung tissue and terminal bronchiolar epithelial cells, and the expression of KL-6 is significantly increased in compensatory hyperplasia of AECs after injuries, which could be caused by connective tissue disease (CTD) [15], viral infections such as COVID-19 [16] and chest radiotherapy[17]. KL-6 can be released from regenerated type II AECs and secreted into the bloodstream through the damaged alveolar basement membrane [15].

As an irAE with high mortality, CIP still lacks effective markers and prediction models. In this study, we aimed to identify risk factors and construct nomograms for evaluating the individual risk of CIP. We enrolled a total of 369 PD-1/PD-L1 monoclonal antibody-treated cancer patients as the training cohort and validation cohort. In the training cohort, variables associated with CIP risk in the differential analysis were included in the binary logistic regression analysis to identify independent risk factors, and a nomogram was established based on the variables. Then, we used a calibration curve [18], receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA) [19] to assess the goodness of fit, accuracy and applicability of the predictive nomogram in the training and validation cohorts.



A total of 245 immunotherapy-treated cancer patients from Shengjing Hospital (as a training cohort) and 124 patients from Liaoning People's Hospital (as a validation cohort) were enrolled in this study between January 1, 2017, and December 31, 2021, and were followed up until February 28, 2022. All of these patients had a definite pathological diagnosis of cancer and were treated with PD1/PD-L1 monoclonal antibodies. The patients with rheumatologic disease, radiation pneumonitis, drug-induced pneumonia and other pre-existing lung abnormalities were excluded from the study. At the same time, other information from these patients, including demographic characteristics (including age, sex, body mass index (BMI), smoking history), the primary and secondary sites of the tumour and prior oncologic therapy, were also collected.

Imaging examination

High Resolution Computed Tomography (HRCT), which was obtained with 1.0–1.5 mm at 10-mm intervals at end inspiration from the lung apex to the base, was performed before and every three months during the PD-1/PD-L1 monoclonal antibody administration and reviewed by two independent radiologists (Liu Yang and Li Hongyi) blinded to the diagnosis and clinical course of the patients. Each radiologist described the phenotypic appearance and severity of CIP according to the CT findings: ground-glass attenuation (GGA), consolidation, traction bronchiectasis or honeycombing [20, 21]. At the same time, the radiologists and clinicans distinguished CIP from diseases such as cancerous lymphangitis, pulmonary infection, and alveolar haemorrhage by clinical, labratory and radiology examination (Fig. 1 shows images from 2 typical CIP patients).

Fig. 1
figure 1

A A 59-year-old male patient with colon cancer and lung metastasis was treated with nivolumab combined with CAPEOX. One month after starting treatment, he developed an irritating dry cough and back pain. HRCT showed thickening of the subpleural lobular septum and multiple reticular changes, which were cured after glucocorticoid administration. B A 66-year-old male patient with liver cancer with intrahepatic metastasis was treated with toripalimab combined with bevacizumab. Dyspnoea occurred after 6 months of treatment. HRCT showed scattered multiple fuzzy patches in both lungs. The application of glucocorticoid and anti-infection treatment did not work, and the patient ultimately died of respiratory failure

Laboratory examination

Before administration of treatment, blood samples were collected and immediately frozen at − 80 °C until used to obtain a series of laboratory values, including acute phase proteins (IL-6 and CRP), T lymphocyte count and serum alveolar epithelial proteins (including SP-A, SP-D and KL-6).

Taking KL-6 as an example, the serum was centrifuged to obtain the supernatant and diluted at 1: 100 in bovine serum albumin (BSA) to prepare the testing sample. A commercial enzyme-linked immunosorbent assay (ELISA) kit (Abebio, AE36646HU) was used, and the detailed protocol was as described previously [22, 23]. Briefly, the test and standard samples were added to antibody-coated 96-well plates and incubated at 37 °C for 2 h. After washing with PBST, the secondary antibody was added, and then the detection solution was added. Finally, the optical density (OD) values were measured with a multimode plate reader, and the concentration of KL-6 was calculated by a standard curve.

Statistical analysis

Continuous variables are expressed as the mean ± SD and were compared by Student's t-test, while categorical variables are presented as counts and percentages and were compared using the chi-square test or Fisher’s exact test. Cut-off values of the continuous variables were determined via ROC analysis using MedCalc software. Difference analysis and binary logistic regression analyses were used to analyse the risk factors, and the risk factors with P < 0.1 in the difference analysis were included in the binary logistic regression analysis. In addition, variance inflation factor (VIF) values were calculated to measure the degree of multicollinearity among the variables; that is, a VIF of > 5 indicates a high correlation of the variables [24], and at the same time, Kolmogorov–Smirnova and Box–Tidwell transformation were employed for normality and linear checks, respectively.

All statistically significant risk factors in the multivariate analysis were used to construct the nomogram using R statistical software. We used calibration curves to measure the agreement between the predicted and actual outcomes. The predictive ability of the nomogram was assessed according to the area under the curve (AUC) of the ROC curve with an AUC closer to 1.0 indicating better results [25]. The overall area under each probable risk threshold was calculated using DCA. In all analyses, P < 0.05 was considered to indicate statistical significance.


Clinicopathological characteristics

During the study period, 245 patients from Shengjing Hospital and 124 patients from Liaoning People's Hospital were recruited for this study. The follow-up time in both groups was less than 2 years, which was slightly longer in the training cohort, and 8.40% of all patients suffered from CIP, most of whom were grades 2–3 and received glucocorticoid therapy. One patient died of interstitial pneumonia complicated by severe pulmonary fungal infection, while the other patients were cured. The basic information of the recruited patients is shown in Table 1. EGFR-TKI therapy in the validation cohort was significantly less common than that in the training cohort, while the other variables were not significantly different between the two cohorts, so we excluded this variable from the following analysis.

Table 1 Clinical characters of training and validation cohort

Difference analysis

As shown in Table 2, we found that smoking history, pleural effusion, number of treatment lines, acute phase proteins (IL-6 and CRP), CD8 + T lymphocyte count and serum alveolar epithelial proteins (including SP-A, SP-D and KL-6) were different between the CIP and non-CIP groups in the training cohort (P < 0.1). Therefore, we included these variables in the logistic regression analyses.

Table 2 Difference analysis of variables in training cohort

All VIF values of the variables were close to 1, indicating no collinearity among the independent variables, and the normality test indicated that the variables were normally distributed, but Box–Tidwell analysis indicated that the relationships between the variables and CIP were nonlinear; therefore, we classified these continuous variables for logistic analysis. A ROC curve and the Youden index (sensitivity + specificity-1) were used to determine the optimal cut-off values of the variables due to the lack of a reference value for cancer patients in the recent literature. When CIP was defined as an endpoint, the optional cut-off values were listed (Table 3), and significant differences in the CIP rate were observed between the patients in the two groups (Fig. 2). By applying these cut-off values, the binary values were divided into two groups, and forest plots revealing the ORs between the two groups are shown in Fig. 3.

Table 3 Optional cut-off values of the variables
Fig. 2
figure 2

Diagnostic value of variables in immunotherapy-administrated patients. ROC of inflammatory markers (A, including IL-6 and CRP), CD8 + T Lymphocyte count (B) and serum alveolar protein (C, including SP-A, SP-D and KL-6)

Fig. 3
figure 3

Forest plot of multivariate regression analysis for ICI-P. Horizontal axis: Odds ratio on a log scale with the reference line, odds ratio (circle) and 95% CI (whiskers)

Binary logistic regression

Binary logistic regression analysis was performed for CIP risk based on clinicopathological features. The results showed that smoking history (OR = 168.56 95%CI 5.03–5643.91, P < 0.01), IL-6 (OR = 25.28 95%CI 1.9–335.89, P = 0.01), CRP (OR = 12.85 95%CI 1.21–137.02, P = 0.03), CD8 + T lymphocyte count (OR = 74.68 95%CI 4.09–1364.72, P < 0.01), SP-A (OR = 46.92 95%CI 2.54–867.77, P = 0.01) and KL-6 (OR = 123.44 95%CI 6.3–2420.57, P < 0.01) were independent influence factors of CIP.

Nomogram construction

These independently associated risk factors were used to form CIP risk estimation nomogram (Fig. 4). The sum of points in the the bottom scales of the nomogram demonstrated the probability of CIP. The nomogram revealed that smoking history was the most influential prognostic factor, closely followed by KL-6 values. In addition, IL-6 values, CD8 + T lymphocyte count and SP-A values also made a moderate contribution to the survival outcome, while CRP value played minor roles.

Fig. 4
figure 4

The establishment of clinical prognostic nomogram models to predict ICI-P risk

Nomogram evaluation

The excellent accuracy of the prediction value of the nomogram was also assessed by the calibration curves, and a preferable consistency between the nomogram-predicted and actual observed values was observed (Fig. 5A). As shown in Fig. 5B, the AUC value of the nomogram for CIP in the training cohort was 0.98 (0.994–1.00), which was higher than the AUC values of any variables alone. Finally, DCA was conducted to assess the clinical utility of the nomogram, and the nomogram showed good consistency in forecasting the CIP risk at all thresholds in the training cohort, implying the good capability of our model (Fig. 5C).

Fig. 5
figure 5

The evaluation of nomogram. A, D The calibration curves for predictions of ICI-P of training cohort (A) and validation cohort (D). The dashed line indicated ideal predictions, the solid line represents actual predictions of nomogram. The closer the distance of two lines, the better the performance of the predictive model. B, E ROC curves for the nomogram of training cohort (B) and validation cohort (E). The AUCs exceed 0.8, which demonstrated that the nomogram could predict the risk of ICI-P. C, F Decision curve analysis for the nomogram of training cohort (C) and validation cohort (F). The black line represents the net benefit at the time when no patients have ICI-P, while the blue line represents the net benefit at the time when all patients have ICI-P; the red line represents a model curve. The area under the three lines demonstrates the clinical usefulness of the nomogram

To further evaluate the nomogram, we assessed its predictive performance in the validation cohort (Fig. 5D–F). The calibration curve in the validation cohort indicated that the model had good discrimination, and the AUC of the nomogram was 0.88 (0.73–0.91). In addition, the DCA illustrated that the nomogram had favourable potential clinical applicability in predicting the CIP risk in the validation cohort, with a 0–0.5 threshold probability. In conclusion, validation of the nomogram showed a good level of agreement with the predictive value.


Biomarkers are usually defined as changed indicators of objectively measured physiological/pathological processes or pharmacological responses to therapeutic interventions [26], which play an important role as a tool for differential diagnosis, prognosis and monitoring disease progression. A nomogram is an intuitive tool for assessing the individual probability of a clinical event based on a statistical predictive model [27], which is also used to predict the risk of adverse events related to tumour treatment [28, 29]. However, few studies have reported irAE-related nomograms, especially for cancers other than lung cancer [30]. In this study, we established a nomogram to predict the risk of CIP according to the baseline clinical characteristics and laboratory values.

The definition of CIP is the occurrence of respiratory symptoms/signs associated with a new emerging infiltration on chest imaging [31]. The causes of CIP include damage to AECs by inflammatory cytokines and activated CD8 + T cells and the recognition of some preexisting antigens (such as ANCA) on AECs by the PD-(L)1 monoclonal antibody. In all kinds of cancers, the overall incidence of CIP varies from 3 to 5% for all grades [32], which is slightly lower than that in our study. The main reasons for include the following: in our study, routine HRCT was used, and more patients were treated with posterior-line therapy.

IL-6 was found in 1973 as a soluble factor that is secreted by T cells and it has been shown to be involved in T-cell activation and the induction of cytokines. CRP was initially found to be a serum protein and it is synthesized in the liver, and its production could be stimulated by IL-6 [33]. IL-6 and CRP are proinflammatory cytokines involved in immune cell recruitment, proliferation and effector functions. IL-6 plays a key role in the systemic inflammatory response, and the IL-6 receptor antagonist tocilizumab showed clinical improvement of a wide variety of irAEs [34]. Previous studies reported that IL-6 and CRP elevation could predict irAEs, especially CIP [35, 36]. We also found that elevated IL-6 and CRP were independent risk factors for CIP, and IL-6 accounted for a larger proportion.

The significant upregulation of activated CD8 + T lymphocytes by ICIs might be the first trigger and it plays a vital role in the occurrence and development of irAEs. The local infiltration of CD8 + T lymphocytes is an important cause of irAEs. Zhou et al. found that the higher the proportion of CD8 + T cells in lung cancer tissue was, the higher the probability of CIP [37]. In addition, the local infiltration of CD8 + T lymphocytes in the skin and hair follicles was also related to immunotherapy-related psoriasis, such as dermatitis [38] and alopecia [39]. Wang et al. also found that pretreatment absolute lymphocyte count was related to an increased risk of irAEs [40]. However, there is still no report on the relationship between lymphocyte taxonomic count and irAEs. In this study, we found elevated CD8 + T lymphocytes in the peripheral blood in CIP patients, suggesting that lymphocyte infiltration may play an important role in CIP.

Radiographic features are currently important determinants for the diagnosis, severity and prognostic assessment of CIP; however, certain serum protein expression levels are also correlated with the severity of CIP and may serve as biomarkers for determining the clinical aspects of CIP. These serum proteins include alveolar epithelial proteins (including surfactant protein and KL-6), chemokines and cytokines (including CCL18, CCL2, CXCL10 and YKL-40) and MMPs and tissue inhibitors of MMPs. In our study, we found that serum alveolar epithelial proteins before treatment could predict CIP risk, but unfortunately, we were not able to study other serum proteins in predicting CIP. A previous study showed that changes in KL-6 levels are correlated with interstitial pneumonia caused by CTD [41]. Therefore, the relationship between serum alveolar epithelial proteins and the efficacy of glucocorticoids in the treatment of CIP needs to be further studied.

Many factors can damage AECs, such as preexisting pulmonary diseases, including chronic obstructive pulmonary disease (COPD), asthma and pleural effusion [32], and some anticancer treatments, such as chemotherapy [42], targeted therapy [43] and radiotherapy [44]. In this study, we found that smoking history and pleural effusion were associated with CIP, while smoking history was an independent risk factor for CIP. Smoking history, which is easily available information, could be used as a guide in clinical practice for irAE prediction.

There are also many shortcomings of this study. First, this study is a dual-centre study from northeastern China, and there may be inclusion bias in the process of recruiting patients. Second, patients with CIP failed to undergo bronchoalveolar lavage or lung biopsy, which helps to clarify the specific lesions of CIP, especially the infiltration of immune cells. Finally, the sample size of this study was small, and no CTLA-4 inhibitors were used by the included patients. We look forward to a large-scale, multicentre prospective study to explore additional irAEs, especially CIP-related markers.


Smoking history, acute phase proteins (IL-6 and CRP), CD8 + T lymphocyte count and serum alveolar epithelial proteins (SP-A and KL-6) are the risk factors of CIP. By combining risk factors, nomograms were constructed for CIP. The models provide an early prediction method for CIP before ICI administration, which facilitates early diagnosis and rational treatment.

Availability of data and materials

All data generated or analyzed in this study (with patient information hidden) can be obtained from the corresponding authors upon reasonable request.



Alveolar epithelial cells


Area under the curve


Body Mass Index


Bovine serum albumin


Immune checkpoint inhibitors pneumonia


Chronic obstructive pulmonary disease


Connective tissue disease


Decision curve analysis


Epidermal growth factor receptor-tyrosine kinase inhibitor


Enzyme linked immunosorbent assay


Ground-glass attenuation


High resolution computed tomography


Immune checkpoint inhibitors


Immunotherapy related adverse events


Krebs von den Lungen-6


Optical density


Programmed Death-(Ligand) 1


Receiver operating characteristic


Surfactant protein A


Variance inflation factor


  1. Xia C, Dong X, Li H, Cao M, Sun D, He S, et al. Cancer statistics in China and United States, 2022: profiles, trends, and determinants. Chin Med J (Engl). 2022;135(5):584–90.

    Article  Google Scholar 

  2. Stucci S, Palmirotta R, Passarelli A, Silvestris E, Argentiero A, Lanotte L, et al. Immune-related adverse events during anticancer immunotherapy: pathogenesis and management. Oncol Lett. 2017;14(5):5671–80.

    PubMed  PubMed Central  Google Scholar 

  3. Zhai X, Zhang J, Tian Y, Li J, Jing W, Guo H, et al. The mechanism and risk factors for immune checkpoint inhibitor pneumonitis in non-small cell lung cancer patients. Cancer Biol Med. 2020;17(3):599–611.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Suresh K, Voong KR, Shankar B, Forde PM, Ettinger DS, Marrone KA, et al. Pneumonitis in non-small cell lung cancer patients receiving immune checkpoint immunotherapy: incidence and risk factors. J Thorac Oncol. 2018;13(12):1930–9.

    Article  CAS  PubMed  Google Scholar 

  5. Kalisz KR, Ramaiya NH, Laukamp KR, Gupta A. Immune checkpoint inhibitor therapy-related pneumonitis: patterns and management. Radiographics. 2019;39(7):1923–37.

    Article  PubMed  Google Scholar 

  6. Naidoo J, Wang X, Woo KM, Iyriboz T, Halpenny D, Cunningham J, et al. Pneumonitis in patients treated with anti-programmed death-1/programmed death ligand 1 therapy. J Clin Oncol. 2017;35(7):709–17.

    Article  CAS  PubMed  Google Scholar 

  7. Elhai M, Hoffmann-Vold AM, Avouac J, Pezet S, Cauvet A, Leblond A, et al. Performance of candidate serum biomarkers for systemic sclerosis-associated interstitial lung disease. Arthritis Rheumatol. 2019;71(6):972–82.

    Article  CAS  PubMed  Google Scholar 

  8. Xue M, Guo Z, Cai C, Sun B, Wang H. Evaluation of the diagnostic efficacies of serological markers KL-6, SP-A, SP-D, CCL2, and CXCL13 in idiopathic interstitial pneumonia. Respiration. 2019;98(6):534–45.

    Article  CAS  PubMed  Google Scholar 

  9. Hasegawa Y, Takahashi M, Ariki S, Saito A, Uehara Y, Takamiya R, et al. Surfactant protein A down-regulates epidermal growth factor receptor by mechanisms different from those of surfactant protein D. J Biol Chem. 2017;292(45):18565–76.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Yamaguchi H, Soda H, Nakamura Y, Takasu M, Tomonaga N, Nakano H, et al. Serum levels of surfactant protein D predict the anti-tumor activity of gefitinib in patients with advanced non-small cell lung cancer. Cancer Chemother Pharmacol. 2011;67(2):331–8.

    Article  CAS  PubMed  Google Scholar 

  11. Ishikawa N, Hattori N, Yokoyama A, Kohno N. Utility of KL-6/MUC1 in the clinical management of interstitial lung diseases. Respir Investig. 2012;50(1):3–13.

    Article  PubMed  Google Scholar 

  12. Tanaka S, Hattori N, Ishikawa N, Shoda H, Takano A, Nishino R, et al. Krebs von den Lungen-6 (KL-6) is a prognostic biomarker in patients with surgically resected nonsmall cell lung cancer. Int J Cancer. 2012;130(2):377–87.

    Article  CAS  PubMed  Google Scholar 

  13. Shoji F, Yamazaki K, Kouso H, Mori R, Takeo S. Predictive impact for postoperative recurrence of preoperative serum Krebs von den Lungen-6 concentration in pathologic stage IA non-small cell lung cancer. Ann Thorac Surg. 2016;101(5):1903–8.

    Article  PubMed  Google Scholar 

  14. Ishikawa N, Hattori N, Yokoyama A, Tanaka S, Nishino R, Yoshioka K, et al. Usefulness of monitoring the circulating Krebs von den Lungen-6 levels to predict the clinical outcome of patients with advanced nonsmall cell lung cancer treated with epidermal growth factor receptor tyrosine kinase inhibitors. Int J Cancer. 2008;122(11):2612–20.

    Article  CAS  PubMed  Google Scholar 

  15. Ma H, Lu J, Song Y, Wang H, Yin S. The value of serum Krebs von den Lungen-6 as a diagnostic marker in connective tissue disease associated with interstitial lung disease. BMC Pulm Med. 2020;20(1):6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Naderi N, Rahimzadeh M. Krebs von den Lungen-6 (KL-6) as a clinical marker for severe COVID-19: a systematic review and meta-analyses. Virology. 2022;566:106–13.

    Article  CAS  PubMed  Google Scholar 

  17. Kleespies RG, Marshall SD, Schuster C, Townsend RJ, Jackson TA, Leclerque A. Genetic and electron-microscopic characterization of Rickettsiella bacteria from the manuka beetle, Pyronota setosa (Coleoptera: Scarabaeidae). J Invertebr Pathol. 2011;107(3):206–11.

    Article  CAS  PubMed  Google Scholar 

  18. Alba AC, Agoritsas T, Walsh M, Hanna S, Iorio A, Devereaux PJ, et al. Discrimination and calibration of clinical prediction models: users’ guides to the medical literature. JAMA. 2017;318(14):1377–84.

    Article  PubMed  Google Scholar 

  19. Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making. 2006;26(6):565–74.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Johkoh T, Fukuoka J, Tanaka T. Rare idiopathic intestinal pneumonias (IIPs) and histologic patterns in new ATS/ERS multidisciplinary classification of the IIPs. EUR J Radiol. 2015;84(3):542–6.

    Article  PubMed  Google Scholar 

  21. Qin H, Xu XP, Zou J, Zhao XJ, Wu HW, Zha QF, et al. Krebs von den Lungen-6 associated with chest high-resolution CT score in evaluation severity of patients with interstitial lung disease. Pulmonology. 2019;25(3):143–8.

    Article  CAS  PubMed  Google Scholar 

  22. Feng CM, Wang XM, Li MD, Xu Z, Hua DX, Cheng JY, et al. Serum interleukin-17 predicts severity and prognosis in patients with community acquired pneumonia: a prospective cohort study. BMC Pulm Med. 2021;21(1):393.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Tamasauskiene L, Sitkauskiene B. Systemic and local cytokine profile and risk factors for persistent allergic airway inflammation in patients sensitised to house dust mite allergens. BMC Pulm Med. 2021;21(1):424.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Bayman EO, Dexter F. Multicollinearity in logistic regression models. Anesth Analg. 2021;133(2):362–5.

    Article  PubMed  Google Scholar 

  25. Janssens A, Martens FK. Reflection on modern methods: revisiting the area under the ROC curve. Int J Epidemiol. 2020;49(4):1397–403.

    Article  PubMed  Google Scholar 

  26. Guiot J, Moermans C, Henket M, Corhay JL, Louis R. Blood biomarkers in idiopathic pulmonary fibrosis. Lung. 2017;195(3):273–80.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Iasonos A, Schrag D, Raj GV, Panageas KS. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol. 2008;26(8):1364–70.

    Article  PubMed  Google Scholar 

  28. Cai G, Yu J, Meng X. Predicting prognosis and adverse events by hematologic markers in patients with locally advanced esophageal squamous cell carcinoma treated with neoadjuvant chemoradiotherapy. Cancer Manag Res. 2020;12:8497–507.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Hirasawa Y, Nakashima J, Sugihara T, Takizawa I, Gondo T, Nakagami Y, et al. Development of a nomogram for predicting severe neutropenia associated with docetaxel-based chemotherapy in patients with castration-resistant prostate cancer. Clin Genitourin Cancer. 2017;15(1):176–81.

    Article  PubMed  Google Scholar 

  30. Chao Y, Zhou J, Hsu S, Ding N, Li J, Zhang Y, et al. Risk factors for immune checkpoint inhibitor-related pneumonitis in non-small cell lung cancer. Transl Lung Cancer Res. 2022;11(2):295–306.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Suresh K, Naidoo J, Lin CT, Danoff S. Immune checkpoint immunotherapy for non-small cell lung cancer: benefits and pulmonary toxicities. Chest. 2018;154(6):1416–23.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Zhang Q, Tang L, Zhou Y, He W, Li W. Immune checkpoint inhibitor-associated pneumonitis in non-small cell lung cancer: current understanding in characteristics, diagnosis, and management. Front Immunol. 2021;12:663986.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Potempa LA, Rajab IM, Olson ME, Hart PC. C-reactive protein and cancer: interpreting the differential bioactivities of its pentameric and monomeric, modified isoforms. Front Immunol. 2021;12:744129.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Stroud CR, Hegde A, Cherry C, Naqash AR, Sharma N, Addepalli S, et al. Tocilizumab for the management of immune mediated adverse events secondary to PD-1 blockade. J Oncol Pharm Pract. 2019;25(3):551–7.

    Article  CAS  PubMed  Google Scholar 

  35. Lin X, Deng H, Yang Y, Wu J, Qiu G, Li S, et al. Peripheral blood biomarkers for early diagnosis, severity, and prognosis of checkpoint inhibitor-related pneumonitis in patients with lung cancer. Front Oncol. 2021;11:698832.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Abolhassani AR, Schuler G, Kirchberger MC, Heinzerling L. C-reactive protein as an early marker of immune-related adverse events. J Cancer Res Clin Oncol. 2019;145(10):2625–31.

    Article  CAS  PubMed  Google Scholar 

  37. Bi J, Qian J, Yang D, Sun L, Lin S, Li Y, et al. Dosimetric risk factors for acute radiation pneumonitis in patients with prior receipt of immune checkpoint inhibitors. Front Immunol. 2021;12:828858.

    Article  PubMed  CAS  Google Scholar 

  38. Tanaka R, Ichimura Y, Kubota N, Saito A, Nakamura Y, Ishitsuka Y, et al. Activation of CD8 T cells accelerates anti-PD-1 antibody-induced psoriasis-like dermatitis through IL-6. Commun Biol. 2020;3(1):571.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Lacouture M, Sibaud V. Toxic side effects of targeted therapies and immunotherapies affecting the skin, oral mucosa, hair, and nails. Am J Clin Dermatol. 2018;19(Suppl 1):31–9.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Shi Y, Liu X, Liu J, Zhang D, Liu X, Yue Y, et al. Correlations between peripheral blood biomarkers and clinical outcomes in advanced non-small cell lung cancer patients who received immunotherapy-based treatments. Transl Lung Cancer Res. 2021;10(12):4477–93.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Zhang T, Shen P, Duan C, Gao L. KL-6 as an immunological biomarker predicts the severity, progression, acute exacerbation, and poor outcomes of interstitial lung disease: a systematic review and meta-analysis. Front Immunol. 2021;12:745233.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Sabat C, Ginestet C, Chassagnon G. Gemcitabine and nab-paclitaxel induced interstitial pneumonia. Diagn Interv Imaging. 2021;102(12):763–4.

    Article  PubMed  Google Scholar 

  43. Nakano K, Seto A, Sasaki T, Shimbashi W, Fukushima H, Yonekawa H, et al. Incidence and risk factors of interstitial lung disease of patients with head and neck cancer treated with cetuximab. Head Neck. 2019;41(8):2574–80.

    Article  PubMed  Google Scholar 

  44. Arroyo-Hernandez M, Maldonado F, Lozano-Ruiz F, Munoz-Montano W, Nunez-Baez M, Arrieta O. Radiation-induced lung injury: current evidence. BMC Pulm Med. 2021;21(1):9.

    Article  PubMed  PubMed Central  Google Scholar 

Download references


We thank all the staff of Shengjing Hospital Oncology Department and Liaoning People's Hospital Oncology Department for their support of our research. In the imaging analysis, we are very grateful for the work of Liu Yang and Li Hongyi.


The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author information

Authors and Affiliations



YW and ZGD contributed to the conception and design of the study. FL and XL performed the experiments and statistical analyses. FL prepared the first draft of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Zhenguang Du.

Ethics declarations

Ethics approval and consent to participate

This research was approved by the Ethics Committee of both Shengjing Hospital of China Medical University and Liaoning People's Hospital, and it was conducted in accordance with the Declaration of Helsinki. All involved patients signed informed consent, and all peripheral blood specimens and information were collected after receiving informed consent.

Consent for publication

Not applicable.

Competing interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, X., Lv, F., Wang, Y. et al. Establishment and validation of nomogram for predicting immuno checkpoint inhibitor related pneumonia. BMC Pulm Med 22, 331 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: