Scholarlink

<p>Human papillomavirus (HPV) infection, particularly with HPV types 16 and 18, is the most prevalent viral carcinogen among head and neck cancers. Notably, HPV-positive (HPV+) status is associated with improved survival and treatment response compared to HPV-negative (HPV−) cancers. However, comprehensive and rigorous analyses of HPV serology remain limited by substantial heterogeneity and numerous outliers. In this study, we analyzed 454 serum samples from 151 patients with squamous cell carcinoma of the head and neck, collected at a single institution between 2007 and 2020. Total IgG antibody titers against the E7 oncoproteins of HPV16 and HPV18 were measured using ELISA (Khanal et al., 2015). The 95% confidence intervals (CIs) for HPV16 E7 and HPV18 E7 were (0.424, 0.500) and (0.251, 0.310) at 15&#xa0;min, and (0.784, 0.952) and (0.419, 0.555) at 30&#xa0;min, respectively. To address outliers and improve classification accuracy, we evaluated several normalization techniques, including Min–Max normalization, general normalization methods, and the logit transformation. We ultimately chose the Min–Max normalization method because it demonstrated superior performance among classifiers. Because we needed to analyze samples in batches rather than simultaneously, we accounted for potential batch effects (plate effects) in our analysis. To minimize these effects, we included negative (blank) control samples on each plate and corrected plate effects using the delta method (i.e., by subtracting values from the blank control) before applying Min–Max normalization. After data transformation, we established optimized thresholds to classify samples as HPV16 E7 positive, HPV18 E7 positive, seronegative (negative for both), or cross-reactive (positive for both). We achieved this classification using unsupervised learning algorithms, specifically K-means and hierarchical clustering, applied to HPV measurements at 15&#xa0;min and 30&#xa0;min, as well as their differences. Our approach yielded a straightforward scoring system with cut-off values of 0.2 on both the X and Y axes, enabling robust distinction between positive and negative samples.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Designing and Classification of Human Papillomavirus (HPV) Signatures in Patients with Head and Neck Cancer

  • 期刊论文

  • 发表日期：

    2026年6月29日

查看全文
  • Bakeerathan Gunaratnam,
  • John A. Strickly,
  • Paul A. Bevins,
  • Dat Thinh Ha,
  • Joongho J. Joh,
  • Rebecca A. Redman,
  • Shesh N. Rai

摘要

Human papillomavirus (HPV) infection, particularly with HPV types 16 and 18, is the most prevalent viral carcinogen among head and neck cancers. Notably, HPV-positive (HPV+) status is associated with improved survival and treatment response compared to HPV-negative (HPV−) cancers. However, comprehensive and rigorous analyses of HPV serology remain limited by substantial heterogeneity and numerous outliers. In this study, we analyzed 454 serum samples from 151 patients with squamous cell carcinoma of the head and neck, collected at a single institution between 2007 and 2020. Total IgG antibody titers against the E7 oncoproteins of HPV16 and HPV18 were measured using ELISA (Khanal et al., 2015). The 95% confidence intervals (CIs) for HPV16 E7 and HPV18 E7 were (0.424, 0.500) and (0.251, 0.310) at 15 min, and (0.784, 0.952) and (0.419, 0.555) at 30 min, respectively. To address outliers and improve classification accuracy, we evaluated several normalization techniques, including Min–Max normalization, general normalization methods, and the logit transformation. We ultimately chose the Min–Max normalization method because it demonstrated superior performance among classifiers. Because we needed to analyze samples in batches rather than simultaneously, we accounted for potential batch effects (plate effects) in our analysis. To minimize these effects, we included negative (blank) control samples on each plate and corrected plate effects using the delta method (i.e., by subtracting values from the blank control) before applying Min–Max normalization. After data transformation, we established optimized thresholds to classify samples as HPV16 E7 positive, HPV18 E7 positive, seronegative (negative for both), or cross-reactive (positive for both). We achieved this classification using unsupervised learning algorithms, specifically K-means and hierarchical clustering, applied to HPV measurements at 15 min and 30 min, as well as their differences. Our approach yielded a straightforward scoring system with cut-off values of 0.2 on both the X and Y axes, enabling robust distinction between positive and negative samples.