Turkthoracj, Turkish Thoracic Journal

Background: Relapse remains a major challenge in the long-term management of pulmonary sarcoidosis, yet predictors of relapse-free survival in patients with early-stage disease are incompletely understood. We investigated the clinical determinants of time to relapse in a biopsy-confirmed cohort of patients with pulmonary sarcoidosis followed for up to 15 years. Methods: This retrospective cohort study included 200 adults (mean age 42.0 ± 11.2 years, 72.0% women) with biopsy-confirmed Stage I–II pulmonary sarcoidosis. The primary outcome was relapse requiring re-initiation or escalation of systemic therapy after remission. Relapse-free survival was evaluated using Kaplan–Meier analysis and compared by the log-rank test. Cox proportional hazards models were used to identify independent predictors of relapse. Results: During long-term follow-up, 36 patients (18.0%) experienced disease relapse. Relapse-free survival was significantly lower among patients who received systemic treatment at baseline than among untreated patients (log-rank p < 0.001). In multivariable Cox analysis, baseline systemic treatment (adjusted hazard ratio [HR] 4.27, 95% confidence interval [CI] 1.39–13.09; p = 0.011) and Scadding Stage II disease (HR 2.58, 95% CI 1.26–5.29; p = 0.010) independently predicted relapse. Increasing age was associated with a lower relapse risk (HR 0.70 per 10-year increase, 95% CI 0.50–1.00; p = 0.047). Baseline FVC and sex were not independently associated with relapse, while extrapulmonary involvement showed a borderline association (HR 2.16, 95% CI 0.97–4.80; p = 0.059). Conclusions: Long-term relapse-free survival in early pulmonary sarcoidosis is primarily determined by baseline disease phenotype. Stage II disease identifies patients at increased relapse risk, whereas the association between systemic treatment and relapse most likely reflects confounding by indication rather than a detrimental treatment effect. Time-to-event analysis provides a clinically relevant framework for long-term risk stratification in early-stage sarcoidosis.

Background: Incidental prostatic adenocarcinoma is frequently identified in transurethral resection of the prostate (TURP) and holmium laser enucleation of the prostate (HoLEP) specimens obtained for the management of benign prostatic hyperplasia. Artificial intelligence (AI)-based algorithms have demonstrated promising performance in prostate biopsy assessment; however, their applicability to TURP and HoLEP materials remains insufficiently studied. We evaluated the standalone performance of an AI algorithm trained on prostate needle biopsies for detecting incidental adenocarcinoma in TURP and HoLEP specimens. Methods: A total of 3,241 whole-slide images from 368 consecutive patients (247 HoLEP and 121 TURP cases) received between May 2023 and December 2025 were retrospectively analyzed. All slides were scanned at ×20 magnification using a Hamamatsu NanoZoomer S360 scanner and processed using a commercially available AI algorithm. The algorithm generated heatmap masks highlighting regions suspicious for malignancy. Conventional histopathological diagnosis served as the reference standard. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and area under the receiver operating characteristic curve (AUC-ROC) were calculated at the case level. The Cochran–Armitage trend test was used to assess the relationship between the number of AI-flagged slides and malignancy rates. Results: Thirty-five cases contained adenocarcinoma, of which 33 were correctly identified by the algorithm, yielding an overall sensitivity of 94.3%. Specificity was 76.0%, PPV 29.2%, NPV 99.2%, and accuracy 77.7%. Sensitivity was 92.3% for HoLEP specimens and 100% for TURP specimens. The case-level AUC-ROC was 0.89. Malignancy rates increased significantly with the number of AI-flagged suspicious slides (0.8%, 13.6%, 18.8%, and 44.4% for 0, 1, 2, and ≥ 3 suspicious slides, respectively; p < 0.001). The two false-negative cases were low-volume Grade Group 1 tumors. Conclusions: A biopsy-trained AI algorithm demonstrated high sensitivity and excellent negative predictive value for detecting incidental prostatic adenocarcinoma in TURP and HoLEP specimens. These findings support its potential role as a screening and workflow-assistance tool in routine uropathology practice.

Amaç: PD-L1 immünohistokimyası (IHC), küçük hücreli dışı akciğer kanseri (KHDAK) olan bireylerde immün kontrol noktası inhibitörü (ICI) tedavisini yönlendiren temel bir prediktif biyobelirteç testidir. Ancak antikor klonlarındaki, derecelendirme sistemlerindeki (Tümör Oranı Skoru (TPS), Kombine Pozitif Skor (CPS), İmmün Hücre skoru (IC)) ve pre-analitik/analitik koşullardaki değişkenlik, özellikle KHDAK'ta küçük biyopsiler ve sitolojik örneklerde yorumlamayı ve tekrarlanabilirliği zorlaştırmaktadır. Bu derlemede, tümör heterojenliği, sitolojik kısıtlılıklar ve yapay zekâ (AI) tabanlı dijital patoloji araçlarının gelişen rolü vurgulanarak, KHDAK’ta PD-L1 testindeki güncel uygulamaları, zorlukları ve gelişmeleri gözden geçirmek amaçlanmıştır. Ayrıca radyomiklerin de dahil olduğu multimodal yaklaşımların doku temelli değerlendirmeyi nasıl tamamlayabileceğini ve ICI tedavisi için hasta seçiminde iyileşme sağlayıp sağlayamayacağını incelemeyi hedefledik. Gereç ve Yöntemler: Ocak 2020 ile Haziran 2025 arasında yayımlanmış çalışmalar taranarak, KHDAK’ta PD-L1 ekspresyonunu doğrulanmış klonlar (22C3, 28-8, SP263, SP142), sitoloji–histoloji uyumu, pre-analitik faktörler ve AI tabanlı PD-L1 skorlama platformları ile değerlendiren çalışmalar üzerine kapsamlı bir literatür taraması yapıldı. Veriler, teknik değişkenlikleri, yorumlayıcı farklılıkları ve ortaya çıkan dijital çözümleri ele alan tematik bir sentez ile analiz edildi. Bulgular: KHDAK’ta PD-L1 ekspresyonu, mekânsal heterojenite ve teknik değişkenliklerden etkilenmekte olup tanısal tutarsızlığa yol açmaktadır. Sitolojik örnekler, sınırlı mimari yapı ve fiksasyon artefaktları nedeniyle benzersiz zorluklar oluşturur. Gözlemciler arası değişkenlik, özellikle %1–49 TPS aralığında en yüksektir. AI destekli algoritmalar ve dijital platformlar, iyileştirilmiş tekrarlanabilirlik (κ 0.74’e kadar), doğruluk (%95’e kadar) ve klinik sonuçlarla olası korelasyon göstermiştir. Lunit SCOPE PD-L1 ve HALO Lung PD-L1 AI gibi ticari AI platformları, %92’ye varan doğruluk oranı elde etmiş ve sınırda sonuçlardaki yanlış sınıflandırma oranlarını %18–30 azaltmıştır. PET tabanlı görüntüleme kullanan radyomikler—SUVmax, metabolik tümör hacmi ve heterojenlik indekslerini içeren—özellikle doku örneklemesinin sınırlı olduğu durumlarda invaziv olmayan bir tamamlayıcı yöntem olarak umut vadetmektedir. Sonuç: Güvenilir PD-L1 testi, klonlara özgü validasyon, standart protokollere uyum ve örnek sınırlılıklarının farkında olunmasını gerektirir. AI tabanlı dijital patoloji ve radyomiklerin entegrasyonu, özellikle belirsiz veya sınırlı örneklerde tanısal doğruluğu artırabilir.

Background: Conventional antifibrotic treatment persistence reflects medication retention but may not fully capture whether patients remain both alive and on their initial therapy. We evaluated alive and on initial treatment (AOIT) as a clinically oriented measure of treatment durability in idiopathic pulmonary fibrosis (IPF) and progressive pulmonary fibrosis (PPF). Methods: This retrospective single-center cohort included adults with IPF or PPF treated with pirfenidone or nintedanib between 2018 and 2026. AOIT was defined as the time from antifibrotic initiation to permanent discontinuation/switching or death while known to remain on initial treatment. Kaplan–Meier and Cox regression analyses were performed. Results: Among 86 patients, 37 (43.0%) experienced AOIT loss, including 26 discontinuation/switch events and 11 deaths. AOIT was 75.1%, 63.6% and 46.9% at 12, 24, and 36 months, respectively; median AOIT was 34.9 months (95% CI 21.3–48.4). AOIT did not differ between nintedanib and pirfenidone (log-rank p = 0.395). First-drug persistence was 80.0%, 71.7%, and 61.7% at the corresponding time points, with a median of 52.2 months (95% CI 31.2–73.2). Malignancy (adjusted HR 5.00, 95% CI 1.91–13.09, p = 0.001) and higher baseline DLCO (adjusted HR 1.67 per 10-percentage-point increase, 95% CI 1.21–2.29; p = 0.002) were associated with AOIT loss. Conclusions: AOIT declined over time and was numerically lower than first-drug persistence, reflecting distinct aspects of longitudinal treatment durability. AOIT may complement medication-centered persistence in real-world antifibrotic treatment assessment.