Search

Introduction: In endoscopic procedures of the nasal sinuses, a critical issue for classification tasks is the ambiguous anatomical representations due to their complex composition. We investigated the potential of multi-label image-based classifications of sinus landmark combinations together with explainability methods for machine learning in an assistance function at the application level. By combining image classification and pixel attribution in a navigation function, we provide the surgeon with label predictions and additional localization cues of important pixels relevant to the model output with regard to the input image. Methods: We used 3500 label annotated video sequences from 30 recorded sinus surgeries to fine-tune a pretrained ResNet50 as the feature extractor and a classification head using binary cross-entropy on one-hot encoded target vectors of landmark classes with a RAdam optimizer over 28-32 epochs. Image augmentation and a focal loss function were added to counter over-fitting. An explainability function used the trained model to produce pixel attribution maps for predicted classes regarding individual input images. These gradient maps were summed over all classes, and pixel values above 0.20 were clustered using weighted k-means based on the gradient value of each pixel coordinate. The resulting cluster centroids were then overlapped into the endoscopic image with the predicted landmark classes. Three surgeons investigated three different overlay scenarios in a validation study. Results: The top-1 predictions reached a mean f1-score of 0.47, with the highest values of 0.71 and the lowest with 0.28. Despite over-fitting mitigation, prediction results largely depended on over- and underrepresented classes. Conclusion: The provided explainability function at the application level showed the strong potential of delivering visual cues from prediction results to surgeons at runtime to support human-machine interaction.

The European Larynx Organ Preservation Study (ELOS; NCT06137378) is a prospective, randomized, open-label, two-armed parallel group controlled, phase II multicenter larynx organ preservation (LOP) trial in locoregionally advanced (LA) stage III, IVA/B head and neck squamous cell carcinoma of the larynx or hypopharynx (LHSCC) amenable for total laryngectomy (TL) with PD-L1 expression within tumor tissue biopsy, calculated as CPS = 1. Induction chemotherapy (IC) with docetaxel and cisplatin (TP) followed by radiation will be compared to TP plus PD-1 inhibition by pembrolizumab (MK-3475; 200 mg i.v. starting day 1 q3w for 17 cycles). After a short induction early response evaluation (ERE) 21 ± 3 days after the first cycle of IC (IC-1), responders achieving endoscopic estimated tumor surface shrinkage (ETSS) =30% will get an additional two cycles of IC followed by intensity-modulated radiotherapy 70-72 Gy (EQD2/a/b = 10) aiming at LOP. Nonresponders (ETSS < 30% or progressing disease) will receive TL and bilateral neck dissection followed by postoperative radiation or chemoradiation as recommended by the clinic's multidisciplinary tumor board. Pembrolizumab treatment will be continued in the intervention arm regardless of ETSS status after IC-1 in both responders and laryngectomized nonresponders, independent of subsequent decisions on adjuvant therapy after TL. [ABSTRACT FROM AUTHOR]

Background/Objectives: Laryngeal squamous cell carcinoma (LSCC) is among most frequent malignancies of the head and neck. Recent oncologic research focusses on advanced rather than on early stages. Thus, we aimed to improve the knowledge concerning prognostic factors and survival in early glottic (GC) and supraglottic cancer (SGC). Methods: We retrospectively investigated patients diagnosed in 2007 to 2020 with stage I or II GC (ICD-10-C32.0) or SGC (ICD-10-C32.1, C32.8 or C32.9). For precise discrimination of GC and SGC, pathology reports about biopsy and definitive excision were closely examined and information on clinical characteristics and risk factors were collected before analyzing patterns of risk factors for overall survival (OS) in multivariate Cox regression analyses (mvCox). Results: The cohort included 220 patients with early GC (n = 183) and SGC (n = 37). The GC patients showed significantly improved 5-year OS compared to SGC patients (83.6% vs. 64.9%; p = 0.004), whereas survival according to UICC stage (I vs. II) was not different (p = 0.177). Surgical resection was superior to definitive radiotherapy (RT) for 5-year OS (p < 0.001). Cumulative tobacco consumption of greater than 10 pack years drastically impaired OS (p = 0.024), especially in patients receiving RT (p < 0.001). Supraglottic localization, smoking, and re-resection after initial R1 status consistently were independent prognostic factors in mvCox. Conclusions: Our cohort of early LSCC patients demonstrates significant negative impact of supraglottic localization, older age, tobacco consumption, poor tumor differentiation, and re-resection on OS. Further research is required as there is still lack of evidence on optimal decision-making and therapeutic strategies. [ABSTRACT FROM AUTHOR]

The documentation of a surgical procedure remains a time-consuming task that surgeons must incorporate into their daily routine. However, since a surgical report should be produced immediately after the operation with all impressions of the procedure in mind, a means of automation assistance should be provided. We, therefore, propose a method that generates surgical reports based on keywords stated during the procedure. Our report generation is based on a sequence-tosequence model that is trained on sentence pairs of two consecutive sentences in a surgical report. The known sentence is augmented with a keyword based on the following surgical action to be documented and is then passed into a language model to generate the next sentence. In this way, the complexity of predicting a vast number of possible surgical report phrasings is reduced to a next sentence prediction task. For the language model, an encoder-decoder structure was used with bidirectional 2-layer Long-Short Term Memory (LSTM) units for both components and an attention layer between input and output sentences. The training data consisted of 50 ear-,nose- and throat surgery (ENT) reports with 1500 sentences. The model training was performed in a k-fold cross-validation study with k = 10 and cross-entropy loss as the objective function. The generated reports were investigated using NIST, ROUGE, and METEOR metrics. Additionally, three medical experts identified the report content regarding plausibility and text errors. The trained models reached an accuracy of 0.82 for the next sentence predictions. The generated reports show consistent sentence structures and keyword correspondence for about 70 % of provided keyword sequences. The NIST, ROUGE, and METEOR metrics reached 0.65, 0.71, and 0.64, respectively. The model underperformed for not yet known keyword sequences and shows signs of overfitting when keyword sequences deviate from the baseline of the training set. Our approach for the keyword-augmented generation of surgical reports shows the potential of reducing the text generation complexity by providing a sequence of anchor words. However, the automated generation of surgical reports remains a difficult task due to individual report phrasings and the high variance in keyword sequences.