Patient-Reported Outcomes of Split-Thickness Skin Grafts for Floor of Mouth Cancer Reconstruction.

INTRODUCTION: Patient-reported outcome measures (PROM) on quality of life (QOL) for early-stage floor of mouth carcinoma (FOM-CA) undergoing surgical resection and split-thickness skin graft (STSG) reconstruction have not been established. We have performed a cross-sectional QOL analysis of such patients to define functional postoperative outcomes. METHODS: Patients with pathologic stage T1/T2 FOM-CA who underwent resection and STSG reconstruction at a tertiary academic cancer center reported outcomes with the University of Washington QOL (v4) questionnaire after at least 6 months since surgery. RESULTS: Twenty-four out of 49 eligible patients completed questionnaires with a mean follow-up of 41 months (range: 6-88). Subsites of tumor involvement/resection included the following: (1) lateral FOM (L-FOM) (n = 17), (2) anterior FOM (A-FOM) (n = 4), and (3) alveolar ridge with FOM, all of whom underwent lateral marginal mandibulectomy (MM-FOM) (n = 3). All patients reported swallowing scores of 70 ("I cannot swallow certain solid foods") or better. Ninety-six percent (23/24) reported speech of 70 ("difficulty saying some words, but I can be understood over the phone") or better. A-FOM patients reported worse chewing than L-FOM patients (mean: 50.0 vs. 85.3; p = 0.01). All 4 A-FOM patients reported a low chewing score of 50 ("I can eat soft solids but cannot chew some foods"). Otherwise, there were no significant differences between subsite groups in swallowing, speech, or taste. CONCLUSION: STSG reconstructions for pathologic T1-T2 FOM-CA appear to result in acceptable PROM QOL outcomes with the exception of A-FOM tumors having worse chewing outcomes.

Negativity and Positivity in the ICU: Exploratory Development of Automated Sentiment Capture in the Electronic Health Record.

OBJECTIVES: To develop proof-of-concept algorithms using alternative approaches to capture provider sentiment in ICU notes. DESIGN: Retrospective observational cohort study. SETTING: The Multiparameter Intelligent Monitoring of Intensive Care III (MIMIC-III) and the University of California, San Francisco (UCSF) deidentified notes databases. PATIENTS: Adult (≥18 yr old) patients admitted to the ICU. MEASUREMENTS AND MAIN RESULTS: We developed two sentiment models: 1) a keywords-based approach using a consensus-based clinical sentiment lexicon comprised of 72 positive and 103 negative phrases, including negations and 2) a Decoding-enhanced Bidirectional Encoder Representations from Transformers with disentangled attention-v3-based deep learning model (keywords-independent) trained on clinical sentiment labels. We applied the models to 198,944 notes across 52,997 ICU admissions in the MIMIC-III database. Analyses were replicated on an external sample of patients admitted to a UCSF ICU from 2018 to 2019. We also labeled sentiment in 1,493 note fragments and compared the predictive accuracy of our tools to three popular sentiment classifiers. Clinical sentiment terms were found in 99% of patient visits across 88% of notes. Our two sentiment tools were substantially more predictive (Spearman correlations of 0.62-0.84, p values < 0.00001) of labeled sentiment compared with general language algorithms (0.28-0.46). CONCLUSION: Our exploratory healthcare-specific sentiment models can more accurately detect positivity and negativity in clinical notes compared with general sentiment tools not designed for clinical usage.

Skin Color Match in Head and Neck Reconstructive Surgery.

OBJECTIVES/HYPOTHESIS: To quantify the degree of color match achieved during microvascular facial reconstruction, and to describe a novel technique for improving reconstructive skin color match. We hypothesize that split-thickness skin grafts (STSG) placed atop de-epithelialized free tissue produces better facial skin color match than free tissue with intact epithelium. STUDY DESIGN: Cross sectional photographic study of reconstructed facial skin color match. METHODS: Sixty-eight adults, who underwent head and neck reconstructive surgery, were divided into six categories based on cutaneous reconstructive technique: cervicofacial flap, radial forearm free flap (RFFF), fibula free flap, anterolateral thigh free flap (ALT), STSG over adiopofascial flap (STAFF), and STSG over myogenous flap (STMF). Averaged color samplings of the reconstructed defect and adjacent normal skin were taken from digital photographs. The color difference was calculated using the delta-E calculation. Blinded expert observers also rated the degree of color match. Nonparametric cohort contrast and correlation statistical analyses were performed. RESULTS: The mean delta-E's and 10-point Likert ratings for the ALT, fibula, RFFF, STAFF, STMF, and cervicofacial flaps were 11.6, 10.0, 7.7, 6.3, 8.8, and 4.7, and 5.1, 6.4, 2.4, 3.2, 2.7, and 1.1, respectively. Likert scale inter-rater correlation was strong, with coefficient = 0.80. CONCLUSIONS: On average, STSG over de-epithelialized myogenous and adipofascial free tissue transfers produced a better color match than the skin paddles of donor sites, with the exception of the radial forearm donor site. Delta-E values obtained from photos correlated well with expert ratings of color match. This reliable technique for quantifying color match may be used in future studies. LEVEL OF EVIDENCE: 3 Laryngoscope, 132:1753-1759, 2022.

Patterns of Nodal Metastases and Predictors of Occult Disease in HPV-Associated Oropharynx Cancer.

OBJECTIVE: For human papilloma virus-associated oropharynx squamous cell carcinoma (HPV+ OPSCC), we evaluated the distribution of neck-level lymph node (LN) metastasis, based on postsurgical histopathology, and the incidence of and risk factors for occult LN metastases, as these patterns need clarification for this newer cancer subset. STUDY DESIGN: Retrospective cohort study. SETTING: National Cancer Database (NCDB). METHODS: We analyzed 2358 patients in the NCDB with HPV+ OPSCC who underwent neck dissection (ND) from 2010 to 2015. Incidence and distribution of LN metastases were calculated for neck levels I to V. Variables associated with occult LN metastasis were assessed by multivariate logistic regression. RESULTS: In therapeutic NDs (n = 1935), the following proportions of positive LNs were found: level I, 9.0% (n = 175); level II, 81.0% (n = 1568); level III, 29.6% (n = 573); level IV, 11.9% (n = 230); and level V, 4.9% (n = 95). In elective NDs (n = 423), occult-positive LNs were found in 35.8% (n = 152), with the following proportions by level: level I, 3.3% (n = 14); level II, 26.9% (n = 114); level III, 8.7% (n = 37); level IV, 4.0% (n = 17); and level V, 0.2% (n = 1). The presence of occult LNs was independently associated with a Charlson-Deyo score of 1 (odds ratio, 2.26; 95% CI, 1.18-4.31; P = .014) and lymphovascular invasion (odds ratio, 5.91; 95% CI, 3.21-11.18; P < .001). Occult LN metastases were not significantly associated with pT classification, primary site, or number of LNs resected. CONCLUSION: For HPV+ OPSCC, occult nodal disease is common. Therapeutic NDs should encompass at least levels II, III, and IV and possibly I, whereas elective NDs could possibly encompass levels II and III. LEVEL OF EVIDENCE: 4.

Temporalis Tendon Transfer/Lengthening Temporalis Myoplasty for Midfacial Static and Dynamic Reanimation After Head and Neck Oncologic Surgery.

Importance: Facial nerve paralysis is an unfortunate, yet relatively common sequela in the treatment of head and neck cancer. Static options for reanimation of the paralyzed midface and oral commissure offer limited quality-of-life improvement and may stretch over time. Reinnervation offers a prolonged recovery and may not be possible among patients with advanced cancer or in salvage situations. Objective: To determine the improvement in resting tone and dynamic excursion gained by use of temporalis tendon transfer (TTT) and lengthening temporalis myoplasty (LTM) for facial nerve reanimation among patients undergoing oncologic head and neck cancer surgery. Design, Setting, and Participants: We present the largest series of primary and secondary TTT operations with prolonged photographic and video-based data with analysis enabled by FaceGram open-source quantitative methodology. This was achieved using preoperative and postoperative digital photographs of each patient at rest and full smile excursion. Using the JAVA-based program Facegram v1.0, measurements were taken of the oral commissure in relation to a line tangent to the inferior border of the lower lip. All operations were performed at the University of California, San Francisco, following oncologic resections between April 2011 and June 2018. Main Outcomes and Measures: Quantitative analysis of facial measurement with Facegram. Results: Of the 32 patients in our series, 19 (59.4%) were male and 13 (40.6%) were female. The average age was 71.3 ± 10 years. There was a mean 4.89 mm (p < 0.01, 95% confidence interval [CI]: 2.78-6.99) increase in height of affected side oral commissure compared with preoperative measurements during full smile excursion, and a 3.05 mm (p < 0.04, 95% CI: 0.22-5.87) increase in horizontal excursion of smile vector. There was also a significant increase in static support at rest, measuring an average of 5.45 mm (p < 0.01, 95% CI: 2.56-8.33) and 2.71 mm (p < 0.04, 95% CI: 0.18-5.25) increase in affected oral commissure vertical and horizontal positions at rest, respectively. The average postoperative follow-up was 11.4 months (95% CI: 5.6-14.11). In addition, when compared with patients treated with postoperative radiation therapy (9 patients), those who received preoperative radiation (15 patients) had a significant improvement in the resting vertical height of the affected side (mean 5.32 mm (p < 0.05, 95% CI: 0.06-10.57)). Conclusions and Relevance: These findings encourage the use of TTT/LTM for facial nerve paralysis in appropriately selected head and neck oncologic patients.

Risk Factors for and Cost Implications of Free Flap Take-backs: A Single Institution Review.

OBJECTIVE/HYPOTHESIS: For patients undergoing microvascular free tissue transfer (MFTT), we evaluated risk factors and financial implications of operating room (OR) take-back procedures. STUDY DESIGN: Retrospective review at a tertiary care center. METHODS: Patients who underwent MFTT for head and neck reconstruction from 2011 to 2018 were identified. We compared hospital length of stay and overall costs associated with OR take-back procedures. Multivariable regression analysis evaluated factors associated with OR take-backs during the same hospitalization. RESULTS: A total of 727 free flaps were reviewed, and 70 OR take-backs (9.6%) were identified. Mean total length of stay (LOS) in the ICU was 3.4 days versus 6.7 days for non-take-back and take-back flaps, respectively (P < .001). Mean total LOS on the regular floor was 6.3 days versus 13.1 days, respectively (P < .001). This resulted in a cost differential of $33,507 (94.3% increase relative to non-take-back flaps). The total cost associated with an OR take-back was $39,786. Hematomas were the most common cause of take-backs and wound dehiscence was associated with the highest costs. On multivariable analysis, higher ASA class (OR, 2.06; 95% CI, 1.11-3.99; P = .026) and shorter ischemia times (OR, 0.52; 95% CI, 0.29-0.95; P = .030) were independently associated with increased risk of take-backs. CONCLUSIONS: OR take-backs infrequently occur but are associated with a significant increase in financial burden when compared to free flap cases not requiring OR take-back. The large majority of the cost differential lies in a substantial increase of ICU and floor LOS for take-back flaps when compared to non-take-back flaps. LEVEL OF EVIDENCE: 4 Laryngoscope, 131:E1821-E1829, 2021.

Patient-Reported Quality of Life After Resection With Primary Closure for Oral Tongue Carcinoma.

OBJECTIVES/HYPOTHESIS: For early-stage oral tongue carcinoma and carcinoma in situ (ESOTCCIS), we evaluated patient-reported quality-of-life (QOL) outcomes following resection with primary closure (R-PC). STUDY DESIGN: Retrospective review at an academic cancer center. METHODS: Thirty-nine ESOTCCIS patients (Tis, T1, T2) who underwent R-PC without radiation completed the University of Washington Quality of Life Questionnaire Version 4 (UW-QOL) at least 6 months since R-PC (mean = 2.39 years; range = 0.5-6.7 years). We compared UW-QOL scores for pain, swallowing, chewing, speech, and taste to established normative population scores. Multivariable regression analysis evaluated factors associated with QOL impairment. RESULTS: ESOTCCIS patients who underwent R-PC in comparison to the normative population reported significantly worse mean speech (87.7 vs. 98, P < .001) and taste (85.6 vs. 95, P = .002) scores and no significant differences in mean pain (91.7 vs. 86, P = .96), swallowing (100 vs. 98, P = .98), chewing (97.4 vs. 94, P = .98) scores. For speech and taste, 59% (23/39) reported no postoperative change from baseline, whereas 41% (16/39) and 35.9% (14/39) reported mild impairment, respectively. Overall, postoperative QOL was reported as good, very good, or outstanding by 87.2% (34/39). Higher American Society of Anesthesiologists class, cT1 compared to CIS, and ventral tongue involvement were independently associated with worse speech. Age < 60 years was independently associated with worse taste. CONCLUSIONS: ESOTCCIS patients who undergo R-PC without radiation can expect long-term swallowing, chewing, and pain to be in the normative range. Although a majority of patients can expect to achieve normative speech and taste outcomes, R-PC carries the risks of mild speech and/or taste impairments. LEVEL OF EVIDENCE: 4 Laryngoscope, 131:312-318, 2021.

Assessing Free Flap Reconstruction Accuracy of the Midface and Orbit Using Computer-Aided Modeling Software.

Importance: Aesthetically pleasing and functionally stable total maxillectomy reconstruction is among the greatest of challenges within head and neck surgery. Maintaining presurgical midface projection, contour, and cheek volume requires high-fidelity restorative control incorporating the provision of stable osseous support to prevent the relatively commonplace complications of lower lid and globe malposition. Objective: To determine the difference in position of the reconstructed infraorbital rim compared with the contralateral normal side, and its effect on globe position, lower eyelid position, orbital volume, and soft tissue contour. Design, Setting, and Participants: This was a case series from 2011 to 2018 in a tertiary academic center. Twelve adult patients underwent free flap reconstruction of the maxilla and inferior orbit by the senior surgeon. Patients with bilateral defects, intact orbital floors, or orbital exenterations were excluded, as were patients without high-resolution postoperative computed tomography (CT) scans. Exposures: Fibular free flaps with orbital floor plates were used in all but one patient, in whom an anterolateral thigh free flap with mesh was used instead. Digital three-dimensional models of the bone and soft tissue were generated from postoperative CT scans using open-source software. Main Outcomes and Measures: Differences in vertical height and anterior projection of the infraorbital rim on the reconstructed side were compared with the normal side. Differences in globe position and lower lid margin were obtained, as were orbital volumes. Differences in soft tissue contour were also examined in six patients who had CT scans at least 6 months after surgery and/or radiation therapy. Results: The absolute mean differences in infraorbital rim height, globe height, lower lid height, and orbital volume were 3.5 mm, 2.5 mm, 3.2 mm, and 2.4 cm3, respectively. The absolute mean difference in soft tissue contour was 10.3 mm for those patients with >6 months follow-up. Differences in infraorbital rim height and globe height were highly correlated (R = 0.77, p = 0.004), as were differences in infraorbital rim height and lower lid height (R = 0.76, p = 0.004). Conclusions and Relevance: Accurate orbitomaxillary free flap reconstruction is possible with precise positioning of the reconstructed infraorbital rim, as this likely affects the position of the lower lid and globe. This methodology of computer-aided modeling with free open-source software may lay the groundwork for future studies seeking to quantify results of facial reconstruction. Level of Evidence: 4.

Predictors of Mortality in HPV-Associated Oropharynx Carcinoma Treated With Surgery Alone.

OBJECTIVE: Survival outcomes for human papillomavirus-associated oropharynx squamous cell carcinoma (HPV + OPSCC) treated with surgery alone are unclear. To increase understanding, we assessed overall survival (OS) outcomes using the national cancer database (NCDB). METHODS: We conducted a retrospective analysis of OS of 736 NCDB HPV + OPSCC patients who underwent surgery alone from 2010 to 2014 using univariate and multivariate analyses and the Kaplan-Meir method. RESULTS: Multivariable analysis found the following independent risk factors for death: American Joint Commission on Cancer (AJCC) 8th edition pathologic stage(p)N2 versus pN0 disease (hazard ratio [HR], 5.5; P = 0.000006), macroscopic extranodal extension (ENE) versus non-ENE (HR, 4.9; P < 0.02), a positive lymph nodes (LN) percentage of ≥10% (HR, 4.2; P = 0.0002), and five or more positive LNs (HR, 4.9; P = 0.00004). Three-year OS was significantly worse for AJCC 8th edition pN2 versus pN0 but not for 7th edition pN2 versus pN0 disease. Five-year OS was significantly worse for positive versus negative surgical margins, AJCC 8th edition stage II versus I, and either microscopic or macroscopic ENE versus non-ENE positive LNs. For 523 (71%) AJCC 8th edition stage I patients and for 283 (38%) patients who were pT1-T2, with negative margins, pN0-N1, with ≤4 pathologic LNs, without ENE, and with >20 LNs removed during neck dissection, the 3-year OS rates were 93% and 95%, respectively, and the 5-year OS rates were 91% and 95%, respectively. CONCLUSION: In the context of the lack of detail and possible inaccuracies found in the NCDB, surgery alone for AJCC 8th edition stage I HPV + OPSCC, particularly pT1-T2, pN0-N1 with ≤4 pathologic LNs, without ENE, and with negative surgical margins has a high OS. LEVEL OF EVIDENCE: 4 Laryngoscope, 130:E423-E435, 2020.

Representing and extracting lung cancer study metadata: study objective and study design.

This paper describes the information retrieval step in Casama (Contextualized Semantic Maps), a project that summarizes and contextualizes current research papers on driver mutations in non-small cell lung cancer. Casama׳s representation of lung cancer studies aims to capture elements that will assist an end-user in retrieving studies and, importantly, judging their strength. This paper focuses on two types of study metadata: study objective and study design. 430 abstracts on EGFR and ALK mutations in lung cancer were annotated manually. Casama׳s support vector machine (SVM) automatically classified the abstracts by study objective with as much as 129% higher F-scores compared to PubMed׳s built-in filters. A second SVM classified the abstracts by epidemiological study design, suggesting strength of evidence at a more granular level than in previous work. The classification results and the top features determined by the classifiers suggest that this scheme would be generalizable to other mutations in lung cancer, as well as studies on driver mutations in other cancer domains.