Identification of patients' smoking status using an explainable AI approach: a Danish electronic health records case study.

BACKGROUND: Smoking is a critical risk factor responsible for over eight million annual deaths worldwide. It is essential to obtain information on smoking habits to advance research and implement preventive measures such as screening of high-risk individuals. In most countries, including Denmark, smoking habits are not systematically recorded and at best documented within unstructured free-text segments of electronic health records (EHRs). This would require researchers and clinicians to manually navigate through extensive amounts of unstructured data, which is one of the main reasons that smoking habits are rarely integrated into larger studies. Our aim is to develop machine learning models to classify patients' smoking status from their EHRs. METHODS: This study proposes an efficient natural language processing (NLP) pipeline capable of classifying patients' smoking status and providing explanations for the decisions. The proposed NLP pipeline comprises four distinct components, which are; (1) considering preprocessing techniques to address abbreviations, punctuation, and other textual irregularities, (2) four cutting-edge feature extraction techniques, i.e. Embedding, BERT, Word2Vec, and Count Vectorizer, employed to extract the optimal features, (3) utilization of a Stacking-based Ensemble (SE) model and a Convolutional Long Short-Term Memory Neural Network (CNN-LSTM) for the identification of smoking status, and (4) application of a local interpretable model-agnostic explanation to explain the decisions rendered by the detection models. The EHRs of 23,132 patients with suspected lung cancer were collected from the Region of Southern Denmark during the period 1/1/2009-31/12/2018. A medical professional annotated the data into 'Smoker' and 'Non-Smoker' with further classifications as 'Active-Smoker', 'Former-Smoker', and 'Never-Smoker'. Subsequently, the annotated dataset was used for the development of binary and multiclass classification models. An extensive comparison was conducted of the detection performance across various model architectures. RESULTS: The results of experimental validation confirm the consistency among the models. However, for binary classification, BERT method with CNN-LSTM architecture outperformed other models by achieving precision, recall, and F1-scores between 97% and 99% for both Never-Smokers and Active-Smokers. In multiclass classification, the Embedding technique with CNN-LSTM architecture yielded the most favorable results in class-specific evaluations, with equal performance measures of 97% for Never-Smoker and measures in the range of 86 to 89% for Active-Smoker and 91-92% for Never-Smoker. CONCLUSION: Our proposed NLP pipeline achieved a high level of classification performance. In addition, we presented the explanation of the decision made by the best performing detection model. Future work will expand the model's capabilities to analyze longer notes and a broader range of categories to maximize its utility in further research and screening applications.

Description of an activity-based enzyme biosensor for lung cancer detection.

BACKGROUND: Lung cancer is associated with the greatest cancer mortality as it typically presents with incurable distributed disease. Biomarkers relevant to risk assessment for the detection of lung cancer continue to be a challenge because they are often not detectable during the asymptomatic curable stage of the disease. A solution to population-scale testing for lung cancer will require a combination of performance, scalability, cost-effectiveness, and simplicity. METHODS: One solution is to measure the activity of serum available enzymes that contribute to the transformation process rather than counting biomarkers. Protease enzymes modify the environment during tumor growth and present an attractive target for detection. An activity based sensor platform sensitive to active protease enzymes is presented. A panel of 18 sensors was used to measure 750 sera samples from participants at increased risk for lung cancer with or without the disease. RESULTS: A machine learning approach is applied to generate algorithms that detect 90% of cancer patients overall with a specificity of 82% including 90% sensitivity in Stage I when disease intervention is most effective and detection more challenging. CONCLUSION: This approach is promising as a scalable, clinically useful platform to help detect patients who have lung cancer using a simple blood sample. The performance and cost profile is being pursued in studies as a platform for population wide screening.

Lung cancer is responsible for more deaths worldwide than all other cancers. It is often detected with the appearance of symptoms when treatment is limited and outcomes for the patient are much worse. While imaging chest scans can detect disease, they are poorly used even in the United States where it is an approved screening method. When cancer is present, protease enzymes are responsible for making space and modifying the lung tissue for the growing tumor. This report describes a panel of 18 sensors that release a fluorescent signal when these enzymes are present in a blood sample. The signal acts like a fingerprint of activity that can be used to identify people with lung cancer. This sensor platform can detect patients with curable lung cancer and could provide a platform for screening very large populations of at-risk individuals.

The Danish lung cancer registry: A nationwide validation study.

BACKGROUND: This study evaluates the validity of the information in the Danish Lung Cancer Registry (DLCR). Since 2000, the DLCR has been a tool for monitoring interventions and outcome of all Danish lung cancer patients with the intent to streamline and improve treatment and survival. The DLCR receives information from the Danish Patient Registries in addition to clinical information from the treating physicians. In the year 2022, more than 50 papers have been published using DLCR as a data source. However, the DLCR has not previously been validated. METHODS: A random sample of 1000 patients diagnosed with non-small cell lung cancer from 2014 to 2016 and recorded in the DLCR were included for validation. Medical records were reviewed and were considered as the "gold standard" to which data listed in the DLCR were compared. RESULTS: Information was retrieved from medical charts for all patients. Agreement on stage at diagnosis was 90.1 % (95 % CI 88.0-91.9) and on date of diagnoses was 93.8 (95 % CI 92.1-93.2). Agreement on smoking status in pack years (+/- 10 pack years) was 91.2 % (95 % CI 88.6-93.2). The positive predictive value of treatment intent was 87.4 (95 % CI 85.1-89.6). CONCLUSION: The data in the DLCR are complete, detailed and accurate. The comparison of data from the DLCR with the medical records revealed overall high validity of the data in the registry.

Introduction of the Danish Lung Nodule Registry: A part of the Danish Lung Cancer Registry.

BACKGROUND: The majority of lung cancer cases are diagnosed late, resulting in poor prognosis and high mortality rates. Early detection and management of lung cancer can improve patient outcomes and reduce mortality rates. Pulmonary nodules are key factors in the early detection of lung cancer, they are common in high-risk populations and require correct classification to determine whether they are benign or malignant. Over the last decade a steep increase in the number of thoracic CT scans has been seen in Denmark, resulting in substantial resources allocated to CT follow-up of incidentally detected pulmonary nodules. The implementation of a nationwide Danish prospective pulmonary nodule registry is to methodically record pulmonary nodules and thereby evaluate the scope of pulmonary nodule follow-up, the nature of the nodules, and the clinical progression of patients with pulmonary nodules. METHODS: A prospective pulmonary nodule registry (Danish Lung Nodule Registry) will be a natural appendix to the Danish Lung Cancer Registry. Three new ICD-10 classification codes will be introduced, defining the type of nodule: /DR91.1/ Solid nodule /DR91.2/ Part-solid nodule; /DR91.3/ Non-solid nodule. Furthermore, an additional letter will describe whether the imaging exam is performed on suspicion of lung cancer (A), or the finding is incidental (B). Registration of the nodules will be performed by the departments of respiratory medicine who manage follow-up of pulmonary nodules. It is estimated that around 7000 nodules will be registered annually. DISCUSSION: The registration of patients in the lung nodule registry complies with current Danish legislation. The registry will be seamlessly integrated with other nationwide Danish registries, including the Danish Lung Cancer Registry, to collect additional patient data and improve the quality and scope of the data acquired. The results from these comprehensive epidemiological studies will be of significant interest and offer valuable research opportunities.

Methylated Cell-Free Tumor DNA in Sputum as a Tool for Diagnosing Lung Cancer-A Systematic Review and Meta-Analysis.

Lung cancer is the leading cause of cancer-related mortality worldwide. Early diagnosis is pivotal for the prognosis. There is a notable overlap between lung cancer and chronic bronchitis, and the potential use of methylated tumor DNA in sputum as a biomarker for lung cancer detection is appealing. This systematic review and meta-analysis followed the PRISMA 2020 statement. A comprehensive search was conducted in Embase, Medline, Web of Science, and the Cochrane Library, using these search strings: Lung cancer, sputum, and methylated tumor DNA. A total of 15 studies met the eligibility criteria. Studies predominantly utilized a case-control design, with sensitivity ranging from 10 to 93% and specificity from 8 to 100%. A meta-analysis of all genes across studies resulted in a summary sensitivity of 54.3% (95% CI 49.4-59.2%) and specificity of 79.7% (95% CI 75.0-83.7%). Notably, two less explored genes (TAC1, SOX17) demonstrated sensitivity levels surpassing 85%. The study's findings highlight substantial variations in the sensitivity and specificity of methylated tumor DNA in sputum for lung cancer detection. Challenges in reproducibility could stem from differences in tumor site, sample acquisition, extraction methods, and methylation measurement techniques. This meta-analysis provides a foundation for prioritizing high-performing genes, calling for a standardization and refinement of methodologies before potential application in clinical trials.

Consequences of Losing Incidental Pulmonary Nodules to Follow-Up: Unmonitored Nodules Progressing to Stage IV Lung Cancer.

INTRODUCTION: Lung cancer is the leading cause of cancer-related death globally. Incidental pulmonary nodules represent a golden opportunity for early diagnosis, which is critical for improving survival rates. This study explores the impact of missed pulmonary nodules on the progression of lung cancer. METHODS: A total of 4,066 stage IV lung cancer cases from 2019 to 2021 in Danish hospitals were investigated to determine whether a chest computed tomography (CT) had been performed within 2 years before diagnosis. CT reports and images were reviewed to identify nodules that had been missed by radiologists or were not appropriately monitored, despite being mentioned by the radiologist, and to assess whether these nodules had progressed to stage IV lung cancer. RESULTS: Among stage IV lung cancer patients, 13.6% had undergone a chest CT scan before their diagnosis; of these, 44.4% had nodules mentioned. Radiologists missed a nodule in 7.6% of cases. In total, 45.3% of nodules were not appropriately monitored. An estimated 2.5% of stage IV cases could have been detected earlier with proper surveillance. CONCLUSION: This study underlines the significance of monitoring pulmonary nodules and proposes strategies for enhancing detection and surveillance. These strategies include centralized monitoring and the implementation of automated registries to prevent gaps in follow-up.

Trends in COPD mortality from 1983 to 2018: protocol for a population-based cohort study in Denmark.

INTRODUCTION: Chronic obstructive pulmonary disease (COPD) is one of the leading causes of death worldwide, which is partly contributed to the increasing prevalence of COPD owning to a demographic shift towards an older population. Conversely, recent studies on COPD mortality that take this demographic shift in age into account find decreasing overall age-standardised COPD mortality rates over time. This decrease in the age-standardised COPD mortality rate is contributed advances in COPD diagnostics and treatment, decreasing smoking prevalence and general advances in medical care particularly in western countries. However, it is unknown if patients with COPD have experienced a comparable relative increase in survival in line with the general population.Hence, there is a need for longitudinal studies comparing trends in mortality in patients with COPD compared with matched non-COPD individuals from the background population. METHODS AND ANALYSES: This is a cohort study with a matched non-COPD comparator cohort. Data are retrieved from the Danish national registers. Data from multiple registries from 1983 to 2018 will be merged on an individual level using the 10-digit Civil Registration numbers that are unique to each citizen in Denmark. Time trends in mortality in patients with COPD compared with the matched comparator cohort will be examined in three study periods: 1983-1993, 1994-2007 and 2008-2018. ETHICS AND DISSEMINATION: The study is entirely based on registry data and ethical approval is not required according to Danish Law and National Ethics Committee Guidelines. The results will be published in peer-reviewed journals and reported at appropriate national and international conferences.

COPD and Smoking Status - It Does Matter: Characteristics and Prognosis of COPD According to Smoking Status.

Background: Chronic obstructive pulmonary disease is a chronic, often progressive disease, which in most patients is caused by tobacco smoking. Our study focuses on differences in COPD-related outcomes between never smokers, former smokers, and current smokers. Methods: A nationwide, population-based cohort study utilizing Danish health registries. Clinical and socioeconomic variables including smoking status, comorbidities, and dyspnea were obtained. Poisson and Cox Regression were used to calculate the impact of these clinical parameters on the risk of moderate and severe exacerbations and mortality during 12 months of follow-up. Results: A total of 49,826 patients ≥40 years of age, with a hospital diagnosis of COPD in 2008-2017, were identified (mean age 69.2 years, 52% females). A total of 2127 (4%) were never smokers, 29,854 (60%) were former smokers and 17,845 (36%) current smokers. Compared to former and current smokers, never smokers reported a lower modified Medical Research Council score and had a milder COPD stage according to the Global Initiative for Chronic Obstructive Lung Disease classification. During follow-up, never smokers had a significantly lower risk of severe exacerbations (hazard ratio 0.87, 95% confidence interval [CI] 0.78-0.97) and a lower rate of death (mortality ratio 0.75, 95% CI 0.70-0.81) compared to patients with a smoking history. Discussion: Our nationwide study showed that COPD in never smokers is characterized by a lower level of dyspnea, milder lung function impairment, and a lower risk of exacerbations and death. At the same time, we found active smokers to have the highest risk. These findings highlight the need for campaigns to prevent smoking and may help general practitioners as well as other health care professionals to motivate patients with COPD to stop smoking.

Tobacco Exposure and Efficacy of Biologic Therapy in Patients With Severe Asthma: A Nationwide Study From the Danish Severe Asthma Register.

BACKGROUND: Randomized trials of biologics in severe, uncontrolled asthma have excluded patients with a cumulative tobacco exposure of more than 10 pack-years. Therefore, our knowledge of the impact of smoking exposure on the clinical effects of biologics in severe asthma remains incomplete. However, because many patients with asthma are current or former smokers, investigating the potential impacts of tobacco exposure on the effects of biologic treatment is clinically important. OBJECTIVE: To investigate the impact of smoking history and tobacco exposure on the effectiveness of biologic therapy in real-life patients with severe asthma. METHODS: We used data from a complete nationwide cohort of patients with severe asthma who were receiving biologics, the Danish Severe Asthma Register. We divided patients according to smoking history and cumulative tobacco exposure and analyzed data at baseline and after 12 months of biologic treatment. RESULTS: A total of 724 bio-naive patients were identified in the Danish Severe Asthma Register, 398 of whom had never been smokers (55%), 316 were previous smokers (44%), and 10 were current smokers (1%). Within the group of current and former smokers, 37% had 1 to 9 pack-years of tobacco exposure, 26% had 10 to 19 pack-years, and 37% had 20 or more pack-years of tobacco exposure. Patients with tobacco exposure had similar reductions in the number of exacerbations, reductions in maintenance oral corticosteroid use, and improvements in asthma symptoms compared with patients with 0 pack-years. CONCLUSION: Former smoking history and lifetime tobacco exposure do not have an impact on the efficacy of biologics in patients with severe asthma.

Introduction of a systematic examination framework for chronic cough: a before-after cohort study in a clinical setting.

Cough is a condition that can be caused by several different mechanisms. There are numerous guidelines for diagnosing the cause of cough, yet the effect of a well-constructed examination framework has not been investigated. At the Department of Internal Medicine, Lillebaelt Hospital, Vejle, a systematic examination framework for diagnosing cough was introduced. Two hundred consecutive patients referred to the pulmonary outpatient clinic with cough were included. The first 100 patients (Group 1) were included before implementation of the examination framework and diagnosed as usual. The next 100 patients (Group 2) were examined using the systematic framework. The primary endpoint was the number of appointments required to establish a diagnosis. A multivariable Poisson regression was performed, adjusting for age, sex, body mass index, pulmonary function (FEV1/FVC), duration of cough, and smoking status. A diagnosis was established within 1-2 visits in 47% in Group 1 compared to 83% in Group 2. When adjusting for confounders, fewer appointments was required to establish a diagnosis in Group 2 (Incidence rate ratio = 0.713 (95% confidence interval: 0.592-0.859), P = 0.000). Using a systematic examination framework for diagnosing cough may reduce the number of appointments required to establish a diagnosis, seemingly without compromising the diagnostic outcome.

Burden of Disease and Productivity Loss in the European Economic Area in Patients Affected by Fibrosing Interstitial Lung Disease.

INTRODUCTION: Progression of fibrosis in interstitial lung diseases (ILD) has been associated with poor prognosis, lower quality of life for patients and caregivers, and higher healthcare costs. This study estimated the burden of disease and productivity loss of progressively fibrosing ILD, focusing on progressive pulmonary fibrosis other than idiopathic pulmonary fibrosis (non-IPF PPF) and systemic sclerosis-associated ILD (SSc-ILD) in the European Economic Area (EEA). METHODS: An economic model was built to estimate the clinical burden of SSc-ILD and non-IPF PPF. The model was based on published data on disease prevalence and disease burden (in terms of comorbidities, exacerbations, and deaths) as well as on productivity loss (in terms of sick days, early retirement, permanent disability, and job loss). Aggregate income loss was obtained by multiplying productivity loss by the median daily income in each country/area of investigation. A sensitivity analysis was performed to test the impact of the variability of the model assumptions. RESULTS: In the whole EEA, a total of 86,794 and 13,221 individuals were estimated to be affected by non-IPF PPF and SSc-ILD, respectively. Estimated annual sick days associated with the diseases were 3,952,604 and 672,172, early retirements were 23,174 and 5341, permanently disabled patients were 41,748 and 4037, and job losses were 19,789 and 2617 for non-IPF PPF and SSc-ILD, respectively. Annual exacerbations were estimated to be 22,401-31,181 and 1259-1753, while deaths were 5791-6171 and 572-638 in non-IPF PPF and SSc-ILD, respectively. The estimated annual aggregate income loss in EEA, accounting for losses due to annual sick days, early retirements, and permanently disabled patients, was €1433 million and €220 million in non-IPF PPF and SSc-ILD, respectively. The productivity loss due to job losses was €194 million and €26 million in non-IPF PPF and SSc-ILD, respectively. The main driver of aggregate income loss variability was the prevalence. CONCLUSION: The impact of non-IPF PPF and SSc-ILD on society is definitely non-negligible. Actions to reduce the burden on our societies are highly needed.

Who are the vulnerable lung cancer patients at risk for not receiving first-line curative or palliative treatment?

BACKGROUND: To identify non-small-cell lung cancer (NSCLC) patients in need of comprehensive support, we examined the association between patient and disease-related factors of vulnerability related to not receiving guideline-recommended treatment. MATERIAL AND METHODS: We identified 14,597 non-small-cell lung cancer (NSCLC) patients with performance status <3 during 2013-2018 in the Danish Lung Cancer Registry. Multivariate logistic regression models were used to estimate Odds Ratios (ORs) and 95% confidence intervals (CIs) for receiving guideline-recommended treatment according to stage, comorbidities, age, performance status, long distance to hospital, cohabitation status, education and alcohol abuse. RESULTS: 21% of stage I-IIIA NSCLC patients did not receive curative treatment while 10% with stage IIIB-IV did not receive any oncological therapy. Factors associated with reduced likelihood of receiving curative treatment included: advanced stage (OR = 0.45; 95% CI = 0.42-0.49), somatic comorbidity (OR = 0.72; 95% CI = 0.63-0.83), age ≥ 80 years (OR = 0.59; 95% CI = 0.55-0.64), performance status = 2 (OR = 0.33; 95% CI = 0.28-0.39) and living alone (OR = 0.79; 95% CI = 0.69-0.90). Results were similar for stage IIIB-IV NSCLC patients, although a statistically significant association was also seen for long distances to the hospital (OR = 0.71; 95% CI = 0.58-0.86). CONCLUSIONS: Several factors are associated with not receiving guideline-recommended NSCLC treatment with age, performance status, comorbidity and stage being most predictive of no treatment receipt. Efforts should be made to develop support for vulnerable lung cancer patients to improve adherence to optimal first-line therapy.

Comorbidity and mortality in systemic sclerosis and matched controls: Impact of interstitial lung disease. A population based cohort study based on health registry data.

OBJECTIVE: This population-based, matched cohort study evaluates the impact of comorbidities on mortality among systemic sclerosis (SSc) patients with and without interstitial lung disease (ILD). METHOD: Patients with a first-time SSc diagnosis between 2002 and 2015 were identified in the Danish National Patient Registry, separated into two cohorts - with ILD (SSc-ILD) and without ILD (non-ILD SSc), and matched 1:4 with controls from the general population on age, sex, residency and marital status. Comorbidity and mortality data were obtained from national registries. The Deyo-Charlson comorbidity score (DCcs) was used for assessment of the burden of comorbidities. RESULTS: 1732 patients with SSc and 6919 controls were included; 258 (14.9%) patients had SSc-ILD. The hazard ratio (HR) for death was 2.8 (95% CI 2.4-3.3) in SSc, and especially increased in SSc-ILD (HR 4.2 (95% CI 3.2-5.4)), males (HR 3.1 95% CI 2.4-4.1) and younger adults (aged 18-40 (HR 6.9, 95% CI 3.4-14.2) and 41-50 (HR 7.7, 95% CI 3.8-15.6)). In non-ILD SSc, mortality increased with increasing DCcs. Cancer was the most frequent cause of death in SSc (24.9% of deaths) and in controls (33.5%), in SSc followed by musculoskeletal and connective tissue diseases (22.7%); the cause of only 0.8% of deaths among controls. CONCLUSION: The high prevalence of comorbidities in SSc had extensive impact on mortality. Mortality was increased in males, in young adults and in SSc-ILD, underlining the excess mortality associated with ILD. These findings emphasise the importance of timely diagnosis and optimal management of organ involvement and comorbidities in SSc.

Methylated Circulating Tumor DNA in Blood as a Tool for Diagnosing Lung Cancer: A Systematic Review and Meta-Analysis.

Lung cancer is the leading cause of cancer-related deaths, and early detection is crucial for improving patient outcomes. Current screening methods using computed tomography have limitations, prompting interest in non-invasive diagnostic tools such as methylated circulating tumor DNA (ctDNA). The PRISMA guidelines for systematic reviews were followed. The electronic databases MEDLINE, Embase, Web of Science, and Cochrane Library were systematically searched for articles. The search string contained three main topics: Lung cancer, blood, and methylated ctDNA. The extraction of data and quality assessment were carried out independently by the reviewers. In total, 33 studies were eligible for inclusion in this systematic review and meta-analysis. The most frequently studied genes were SHOX2, RASSF1A, and APC. The sensitivity and specificity of methylated ctDNA varied across studies, with a summary sensitivity estimate of 46.9% and a summary specificity estimate of 92.9%. The area under the hierarchical summary receiver operating characteristics curve was 0.81. The included studies were generally of acceptable quality, although they lacked information in certain areas. The risk of publication bias was not significant. Based on the findings, methylated ctDNA in blood shows potential as a rule-in tool for lung cancer diagnosis but requires further research, possibly in combination with other biomarkers.

Diagnosis and treatment of lung cancer in Denmark during the COVID-19 pandemic.

BACKGROUND: We examined the number of lung cancers diagnosed, the quality of care and the socio-economic and clinical characteristics among patients with lung cancer during the COVID-19 pandemic compared to previous years. METHODS: We included all patients ≥ 18 years old diagnosed with lung cancer from 01 January 2018 to 31 August 2021 as registered in the Danish Lung Cancer Registry. Using a generalised linear model, we estimated prevalence ratios (PR) and 95% confidence intervals (CI) of the associations between the pandemic and socioeconomic and clinical factors, and indicators of quality. RESULTS: We included 18,113 patients with lung cancer (82.0% non-small cell lung cancer (NSCLC)), which was similar to the preceding years, although a decline in NSCLC cases occurred during the first lockdown period in 2020. No difference in distribution of income or educational level was observed. No difference was observed in the quality of treatment - as measured by curative intent, proportion of patients resected or who died within 90 days of diagnosis. CONCLUSION: Using nationwide population-based data, our study reassuringly shows no adverse effects of the COVID-19 pandemic on the diagnosis, socio-economic characteristics nor quality of treatment of lung cancer, as compared to the preceding years.

NK cell activity and methylated HOXA9 ctDNA as prognostic biomarkers in patients with non-small cell lung cancer treated with PD-1/PD-L1 inhibitors.

BACKGROUND: PD-1/PD-L1 inhibitors have improved survival for patients with non-small cell lung cancer (NSCLC). We evaluated natural killer cell activity (NKA) and methylated HOXA9 circulating tumor DNA (ctDNA) as prognostic biomarkers in NSCLC patients treated with PD-1/PD-L1 inhibitors. METHODS: Plasma was prospectively collected from 71 NSCLC patients before treatment with PD-1/PD-L1 inhibitors and before cycles 2-4. We used the NK Vue® assay to measure the level of interferon gamma (IFNγ) as a surrogate for NKA. Methylated HOXA9 was measured by droplet digital PCR. RESULTS: A score combining NKA and ctDNA status measured after one treatment cycle had a strong prognostic impact. Group 1 had IFNγ < 250 pg/ml and detectable ctDNA (n = 27), group 2 consisted of patients with either low levels of IFNγ and undetectable ctDNA or high levels of IFNγ and detectable ctDNA (n = 29), group 3 had IFNγ ≥250 pg/ml and undetectable ctDNA (n = 15). Median OS was 221 days (95% CI 121-539 days), 419 days (95% CI 235-650 days), and 1158 days (95% CI 250 days-not reached), respectively (P = 0.002). Group 1 had a poor prognosis with a hazard ratio of 5.560 (95% CI 2.359-13.101, n = 71, P < 0.001) adjusting for PD-L1 status, histology, and performance status. CONCLUSIONS: Combining NKA and ctDNA status after one cycle of treatment was prognostic in patients with NSCLC treated with PD-1/PD-L1 inhibitors.

New insights in the genetic variant spectrum of SLC34A2 in pulmonary alveolar microlithiasis; a systematic review.

Pulmonary alveolar microlithiasis (PAM) is a rare autosomal recessive lung disease caused by variants in the SLC34A2 gene encoding the sodium-dependent phosphate transport protein 2B, NaPi-2b. PAM is characterized by deposition of calcium phosphate crystals in the alveoli. Onset and clinical course vary considerably; some patients remain asymptomatic while others develop severe respiratory failure with a significant symptom burden and compromised survival. It is likely that PAM is under-reported due to lack of recognition, misdiagnosis, and mild clinical presentation. Most patients are genetically uncharacterized as the diagnostic confirmation of PAM has traditionally not included a genetic analysis. Genetic testing may in the future be the preferred tool for diagnostics instead of invasive methods. This systematic review aims to provide an overview of the growing knowledge of PAM genetics. Rare variants in SLC34A2 are found in almost all genetically tested patients. So far, 34 allelic variants have been identified in at least 68 patients. A majority of these are present in the homozygous state; however, a few are found in the compound heterozygous form. Most of the allelic variants involve only a single nucleotide. Half of the variants are either nonsense or frameshifts, resulting in premature termination of the protein or decay of the mRNA. There is currently no cure for PAM, and the only effective treatment is lung transplantation. Management is mainly symptomatic, but an improved understanding of the underlying pathophysiology will hopefully result in development of targeted treatment options. More standardized data on PAM patients, including a genetic diagnosis covering larger international populations, would support the design and implementation of clinical studies to the benefit of patients. Further genetic characterization and understanding of how the molecular changes influence disease phenotype will hopefully allow earlier diagnosis and treatment of the disease in the future.

Fatigue in Patients with Sarcoidosis in Denmark.

INTRODUCTION: Sarcoidosis is a systemic granulomatous disease of unknown etiology, primarily affecting the lungs and thoracic lymph nodes. Fatigue is a frequent and disabling symptom in sarcoidosis with a significant impact on quality of life. In Denmark, the incidence of sarcoidosis has increased; however, the extent and risk factors of fatigue have not been investigated and no recent reports of the characteristics of patients with sarcoidosis in Denmark exist. AIM: To assess the frequency of fatigue in patients with sarcoidosis in Denmark at diagnosis and to investigate if fatigue correlated with relevant disease parameters. Moreover, to characterize patients with sarcoidosis in Denmark at time of diagnosis. METHODS: Data were collected in 150 patients with recently diagnosed sarcoidosis. Fatigue was measured using the Fatigue Assessment Scale (FAS). Patients with fatigue were compared to non-fatigue patients regarding clinical parameters. RESULTS: FAS was completed by 145 of 150 patients. Fifty-one percent reported significant fatigue. Mean FAS score was 23.6 and 51% had a FAS score ≥ 22. Fatigue in 89 incident patients did not correlate significantly with demographics, physiological, or clinical parameters. Fifty-nine percent were males. Mean age was 47 years; mean values (% predicted) for pulmonary function tests were normal and 71% at Scadding stage 0-I. CONCLUSION: In Denmark, fatigue was frequent in patients with sarcoidosis. The majority of patients had mild disease at diagnosis and were older but lower at Scadding stage at diagnosis compared to previous cohorts.

Assessment of circulating biomarkers for detection of lung cancer in a high-risk cohort.

BACKGROUND: There is an urgent need for early detection of lung cancer. Screening with low-dose computed tomography (LDCT) is now implemented in the US. Supplementary use of a lung cancer biomarker with high specificity is desirable. OBJECTIVE: To assess the diagnostic properties of a biomarker panel consisting of cytokeratin 19 fragment (CYFRA 21-1), carcinoembryonic antigen (CEA) and cancer antigen 125 (CA125). METHODS: A cohort of 250 high-risk patients was investigated on suspicion of lung cancer. Ahead of diagnostic work-up, blood samples taken. Cross-validated prediction models were computed to assess lung cancer detection properties. RESULTS: In total 32% (79/250) of patients were diagnosed with lung cancer. Area under the curve (AUC) for the three biomarkers was of 0.795, with sensitivity/specificity of 57%/93% and negative predictive value of 83%. When combining the biomarkers with US screening criteria, the AUC was 0.809, while applying only US screening criteria on the cohort, yielded an AUC of 0.62. The ability of the biomarkers to detect stage I-II lung cancer was substantially lower; AUC 0.54. CONCLUSIONS: In a high-risk cohort, the detection properties of the three biomarkers were acceptable compared to current LDCT screening criteria. However, the ability to detect early stage lung cancer was low.