Background

The purpose of this study was to evaluate the predictive value of positron emission tomography (PET)-CT in identifying occult nodal metastasis in clinically and radiographically N0 patients with recurrent laryngeal cancer undergoing salvage laryngectomy.

Methods

Retrospective review of 46 clinically and radiographically N0 patients with recurrent laryngeal cancer who underwent a PET-CT examination before salvage laryngectomy with neck dissection from January 1, 2002, to December 31, 2014, was performed.

Results

Two patients (16.7%) had true-positive PET-CT results, whereas 10 patients (83.3%) had false-negative scans, 1 patient (2.9%) had a false-positive result and 33 patients (97.1%) had a true-negative PET-CT. The sensitivity of PET-CT was 16.7% (95% confidence interval [CI], 3.5% to 46.0%) with a specificity of 97.1% (95% CI, 83.8% to 99.9%), positive predictive value (PPV) of 66.7% (95% CI, 20.2% to 94.4%), and negative predictive value (NPV) of 76.7% (95% CI, 62.1% to 87.0%).

Conclusion

PET-CT has poor sensitivity and NPV making PET-CT an imperfect predictor of nodal disease in recurrent laryngeal cancer. © 2017 Wiley Periodicals, Inc. Head Neck 39: 980-987, 2017.

Objectives

To report our institutional experience, management, and outcomes of cutaneous periauricular squamous cell carcinoma (SCC).

Study design

Retrospective chart review.

Setting

Tertiary academic center.

Subjects

Patients undergoing treatment of cutaneous periauricular SCC from 2000 to 2016.

Results

A total of 112 patients had a median follow-up of 24.5 months, a mean ± SD age of 75.7 ± 10.6 years, and a strong male predominance (93.8%). Site distribution shows 87 (77.7%) auricular, 26 (23.2%) preauricular, and 10 (8.8%) postauricular lesions. Of auricular lesions, tumors involved the tragus (n = 3, 3.4%), helix/antihelix (n = 47, 54.0%), conchal bowl (n = 31, 35.6%), external auditory canal (n = 18, 16.1%), and lobule (n = 3, 3.4%). Most patients presented at stage I (52.7%) versus stages II (28.6%), III (6.3%), and IV (12.5%). Patients were largely treated surgically with primary tumor resection ranging from wide local excision to lateral temporal bone resection (± parotidectomy and neck dissection), with 17.0% and 5.4% receiving adjuvant radiation and chemoradiation, respectively. Metastatic spread was seen to the parotid (25.9%) and neck (26.8%), with most common cervical spread to level II. Overall survival, disease-specific survival, and disease-free survival at 3 years were 62%, 89%, and 56%, respectively. Nodal disease was associated with worse disease-specific survival (P < .001) and disease-free survival (P = .042). Pre- and postauricular sites were associated with worse overall survival (P = .007) relative to auricular sites.

Conclusion

Among cutaneous SCC, periauricular subsites pose treatment challenges related to surrounding anatomy and represent a unique tumor population. The reported propensity toward recurrence and patterns of metastasis may better guide treatment of aggressive tumors to include regional nodal dissection.

PURPOSE: Report relevance of molecular groups to clinicopathologic features, germline SMARCB1/SMARCA4 alterations (GLA), and survival of children with atypical teratoid rhabdoid tumor (ATRT) treated in two multi-institutional clinical trials. MATERIALS AND METHODS: Seventy-four participants with newly diagnosed ATRT were treated in two trials: infants (SJYC07: age < 3 years; n = 52) and children (SJMB03: age 3-21 years; n = 22), using surgery, conventional chemotherapy (infants), or dose-dense chemotherapy with autologous stem cell rescue (children), and age- and risk-adapted radiotherapy [focal (infants) and craniospinal (CSI; children)]. Molecular groups ATRT-MYC (MYC), ATRT-SHH (SHH), and ATRT-TYR (TYR) were determined from tumor DNA methylation profiles. RESULTS: Twenty-four participants (32%) were alive at time of analysis at a median follow-up of 8.4 years (range, 3.1-14.1 years). Methylation profiling classified 64 ATRTs as TYR (n = 21), SHH (n = 30), and MYC (n = 13), SHH group being associated with metastatic disease. Among infants, TYR group had the best overall survival (OS; P = 0.02). However, outcomes did not differ by molecular groups among infants with nonmetastatic (M0) disease. Children with M0 disease and <1.5 cm2 residual tumor had a 5-year progression-free survival (PFS) of 72.7 ± 12.7% and OS of 81.8 ± 11%. Infants with M0 disease had a 5-year PFS of 39.1 ± 11.5% and OS of 51.8 ± 12%. Those with metastases fared poorly [5-year OS 25 ± 12.5% (children) and 0% (infants)]. SMARCB1 GLAs were not associated with PFS. CONCLUSIONS: Among infants, those with ATRT-TYR had the best OS. ATRT-SHH was associated with metastases and consequently with inferior outcomes. Children with nonmetastatic ATRT benefit from postoperative CSI and adjuvant chemotherapy.

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.