Classificatory updates in verrucous and cuniculatum carcinomas:Insights from the 5^(th) edition of WHO-IARC head and neck tumor classification

The International Agency for Research on Cancer(IARC)and World Health Organization(WHO)collaboratively produce the'WHO Blue Books'essential tools standardizing the diagnostic process for human cancers.Regular updates in this classification accommodate emerging molecular discoveries,advances in immunohistochemical techniques,and evolving clinical insights.The 5th edition of the WHO/IARC classification of head and neck tumors refines the'Oral Cavity and Mobile Tongue'chapter,including sections for non-neoplastic lesions,epithelial tumors,and tumors of uncertain histogenesis.Notably,the epithelial tumors section is rearranged by tumor behavior,starting with benign squamous papillomas and progressing through potentially malignant oral disorders to oral squamous cell carcinoma(OSCC).The section on OSCC reflects recent information on epidemiology,pathogenesis,and histological prognostic factors.Noteworthy is the specific categorization of verrucous carcinoma(VC)and carcinoma cuniculatum(CC),both associated with the oral cavity and distinct in clinical and histologic characteristics.This classification adjustment emphasizes the oral cavity as their predominant site in the head and neck.Designating specific sections for VC and CC aims to provide comprehensive insights into these unique subtypes,elucidating their clinical features,distinct histological characteristics,prevalence,significance,and clinical relevance.By categorizing these subtypes into specific sections,the 5th edition of the WHO classification aims to provide a more nuanced and detailed account,enhancing our understanding of these specific variants within the broader spectrum of head and neck tumors.

An Overview and Comparative Evaluation of Head and Neck Cancer Risk Factors in India and Australia

Aim: An overview and comparative evaluation of head and neck cancer risk factors in India and Australia. Method: In this review we included articles with information on head and neck cancer risk factors and its association based on: smoking tobacco, alcohol, smokeless tobacco, betel nut and areca nut chewing, viral infection like HPV, dental hygiene, diet, family history, socioeconomic status, other heavy metals and systemic conditions. Articles with clinical features, diagnosis, treatment and prognosis were excluded. Results: Head and neck cancer in India has different demographic, risk factors, dietary habits, personal and family history. Oral cancer is more common amongst all head and neck squamous cell cancers in males. This is mainly attributed with consumption of a variety of smokeless tobacco, smoking, alcohol, poverty, illiteracy, cultural, advanced stage at presentation;lack of good treatment infrastructure creates main challenge to India as compared to Australia. Conclusion: The knowledge about risk factors for HNC in public health education for general population supports health promotion and tobacco prevention, which is the main aim of the programs started by the government, as head and neck cancers are potentially preventable.

Signature Based on Six Autophagy-related Genes to Predict Prognosis of Head and Neck Squamous Cell Carcinoma

Objective The present study aimed to develop an autophagy-related gene prognostic prediction model to provide survival risk prediction for head and neck squamous cell carcinoma(HNSCC)patients.Methods The K-mean cluster analysis was performed on HNSCC samples based on the expression values of 210 autophagy-related genes for candidate signature gene selection.LASSO Cox regression analysis was generated using the potential genes and the risk score was calculated from the prognosis model.The risk score was processed as an independent prognostic indicator to construct the nomogram model.The immune status including immune cell infiltration ratio and checkpoints of patients with HNSCC in high-and low-risk groups was also explored.Results LASSO Cox regression analysis was performed on the selected autophagy-related genes.According to the lambda value corresponding to the number of different genes in the LASSO Cox analysis,six genes(GABARAPL2,SAR1A,ST13,GAPDH,FADD and LAMP1)were finally chosen.The risk score based on the genes was generated,which was an independent prognostic marker for HNSCC.The prognostic prediction model(nomogram)was further optimized by the independent prognostic factors(risk score),which can better predict the prognosis and survival of patients.With the risk score and prognosis model,eight types of immune cells and six key immune checkpoints(CTLA4,PD1,IDO1,TDO2,LAG3,TIGIT)displayed expression specificity.Conclusion This study identified several potential prognostic biomarkers and established an autophagy-related prognostic prediction model for HNSCC,which provides a valuable reference for future clinical research.

How Safe Are Reduced Doses per Fraction in Target Volumes of 2^ndto 4^thOrder in the Simultaneous Integrated Boost Irradiation Technique in Head and Neck Carcinoma Patients?

Aim: The simultaneous irradiation of target volumes of different total dose levels using intensity modulated radiotherapy leads to reduced doses per fraction and longer treatment times in target volumes of 2nd?to 4th?order. Does the thereby caused reduced biological effectiveness induce an increased recurrence risk? The current work deals with the problem of recurrences of patients with head and neck carcinomas treated either with an intensitiy (IMRT) or with a volumetric modulated (VMAT) irradiation technique. Methods: From October 2002 to September 2014, 699 patients with carcinomas of the head and neck were irradiated using IMRT or VMAT. The median follow up of the patients was 21.9 months (2 to 145 months). Primary tumor regions (1st?order target volume) of 565 patients were treated with doses per fraction of 2 Gy. Accordingly, further 133 target volumes of the primary tumor received reduced doses per fraction. In 1 patient, the lymphatic drainage was treated solely without irradiation of the primary region. For the lympatic drainage, 854 1st?order target volumes were treated with a dose per fraction of 2 Gy. Reduced doses per fraction were applied to further 1780 target volumes. Results: 54 of 699 patients developed a recurrence in the primary tumor region after radio-(chemo) therapy, 4 patients developed a recurrence of the primary tumor and a unilateral recurrence of the lymphatic drainage, 2 patients a recurrence of the primary tumor and a bilateral lymph node recurrence. 18 patients showed an isolated unilateral recurrence and additionally 2 patients a bilateral recurrence of the lymphatic drainage. 619 patients stayed recurrence free. In primary tumor regions, receiving a dose per fraction of 2 Gy 55 patients (9.7%) developed a recurrence, whereas in target volumes receiving a reduced dose per fraction 5 patients (3.8%) developed a recurrence (p < 0.001). In lympatic drainage target volumes receiving a dose per fraction of 2 Gy, 25 target volumes (2.9%) developed a recurrence, whereas in target volumes receiving a reduced dose per fraction, 5 patients (0.3%) developed a recurrence (p = 0.001). Conclusion: The recurrence risk in target volumes of 2nd?to 4th?order was not increased due to reduced doses per fraction deposited by means of a simultaneous integrated boost technique. Therefore, the simultaneous irradiation of target volumes with different dose levels is safely applicable within one treatment plan.

MultiTrans:Multi-scale feature fusion transformer with transfer learning strategy for multiple organs segmentation of head and neck CT images

Radiotherapy with precise segmentation of head and neck organs at risk(OARs)is one of the important treatment methods for head and neck cancer.In routine clinical practice,OARs are manually segmented by doctors to avoid irreversible adverse reactions caused by radiotherapy,which is time-consuming and laborious.To assist doctors in OARs segmentation,a MultiTrans framework with a multi-scale feature fusion module was proposed in this paper.In the multi-scale feature fusion module,the original image and the feature map of CNN were fused together to form a compound feature map for more complete high-resolution global information.In addition,the global information was also fully utilized in MultiTrans by using the feature map restored from the compound feature map in the skip connection.The multi-scale interactive high-resolution information can make full use of medical image information and obtain features more comprehensively,thus improve the segmentation accuracy.Experiments showed that MultiTrans had an average Dice score coefficient(DSC)of 74.01%in all organs,effectively improved segmentation accuracy.In addition,we proposed a transfer learning strategy for small organs by transferring the weight parameters of organs with a large amount of data to organs with a small amount of data to speed up the convergence of MultiTrans and reduce the demand for data volume in the MultiTrans.With this strategy,the average DSC of small organs was obviously increased,making the segmentation of small organs more accurate.The proposed framework and transfer learning strategy have the potential of assisting doctors in OARs delineation.

38例头颈部髓外浆细胞瘤临床分析

目的探讨头颈部髓外浆细胞瘤(head and neck extramedullary plasmacytoma,HNEMP)的临床特征、危险因素、治疗及预后。方法回顾性分析2008年9月~2022年1月间北京同仁医院收治的38例新诊断且未治疗的HNEMP患者的临床资料,分析影响生存的因素及其预后。结果38例患者中,病变呈现鼻颅眶沟通性者5例(13.2%),肿瘤最大直径≥5 cm者8例(21.1%)。仅接受单纯手术切除的患者17例(44.7%),术后给予放疗者13例(34.2%),化疗者8例(21.1%),接受放化疗联合治疗者5例(13.2%)。治疗后评价疗效,完全缓解32例,部分缓解3例,疾病稳定1例,疾病进展2例。中位随访86个月,5年的无进展生存率和总生存率分别92.1%和94.7%。其中,颅鼻眶沟通病变患者中位无进展生存期仅13个月,5年的无进展生存率和总生存率为40%和60%,而其余组均为100%。此外,病变范围≥5 cm和累及鼻颅眶沟通是影响患者无进展生存率和总生存率的不良预后因素。结论HNEMP发病率低,治疗上以手术联合放疗为主,预后相对较好,颅鼻眶沟通病变及高肿瘤负荷为不良预后因素。

Monaco计划系统中鼻咽癌患者颈部皮肤受量的剂量学研究

目的探索在Monaco计划系统中制定鼻咽癌放疗计划时,如何降低鼻咽癌患者颈部皮肤受照剂量。方法随机选取我院已完成治疗的中晚期鼻咽癌患者计划15例,将放疗靶区上下各1 cm范围内由系统自动生成的皮肤(Skin1)内缩3 mm形成新的皮肤轮廓(Skin2),并由二者通过减法运算得到一个厚度为3 mm的皮肤层环形结构(Skin3)。在Monaco计划系统中按处方要求重新制定放疗计划时,增加Parallel函数对该环形结构进行体积剂量的调控,并对两组计划的数据进行对比分析。结果两组计划的原发部位PGTVnx D_(95)、颈部淋巴结PGTVnd D_(95),高危淋巴引流区PCTV1 D_(95)的靶区剂量分布以及危及器官受照剂量相比较,差异均无统计学意义(P>0.05);两组预防照射区PCTV2 D_(95)和机器输出跳数(MU)相比较,差异有统计学意义(P<0.05),单独调控后机器的输出(MU)会增加6.15%;两组PCTV2的剂量分布适形度指数(CI)和皮肤受照剂量差异有统计学意义(P<0.05),单独调控后PCTV2 CI增加6.41%且皮肤受量下降明显,其中Skin3V 50下降71.56%。结论在Monaco计划系统中对鼻咽癌患者颈部皮肤进行剂量调控,虽然会在一定程度上引起机器输出的增加,但是会对靶区剂量分布进行再调整,剂量约束更加优化,皮肤受量也能得到很好的调控,这对于降低患者皮肤受量提高放疗质量能够起到相应的促进作用。

Risk factors for ventilator-associated pneumonia among patients undergoing major oncological surgery for head and neck cancer

Patients undergoing major oncological surgery for head and neck cancer （SHNC） have a particularly high risk of nosocomial infections. We aimed to identify risk factors for ventilator-associated pneumonia （VAP） in patients undergoing SHNC. The study included 465 patients who underwent SHNC between June 2011 and June 2014. The rate of VAP, risk factors for VAP, and biological aspects of VAP were retrospectively evaluated. The incidence of VAP was 19.6~0 （n = 95） in patients who required more than 48 h of mechanical ventilation. Staphylococcus （37.7%）, Enterobacteriaceae （32.1%）, Pseudomonas （20.8%）, and Haemophilus （16.9~） were the major bacterial species that caused VAP. The independent risk factors for VAP were advanced age, current smoking status, chronic obstructive pulmonary disease, and a higher simplified acute physiology score system II upon admission. Tracheostomy was an independent protective factor for VAP. The median length of stay in the ICU for patients who did or did not develop VAP was 8.0 and 6.5 days, respectively （P = 0.006）. Mortality among patients who did or did not develop VAP was 16.8% and 8.4%, respectively （P 〈 0.001）. The potential economic impact of VAP was high because of the significantly extended duration of ventilation. A predictive regression model was developed with a sensitivity of 95.3% and a specificity of 69.4%. VAP is common in patients who are undergoing SHNC and who require more than 48 h of mechanical ventilation. Therefore, innovative preventive measures should be developed and applied in this high-risk population.

头颈部恶性肿瘤根治术后下肢深静脉血栓形成的风险预测模型构建

目的构建头颈部恶性肿瘤根治术后下肢深静脉血栓形成(lower extremity deep venous thrombosis,LDVT)的风险预测模型并探讨其临床价值。方法回顾性纳入2017年1月—2022年12月于徐州市中心医院口腔颌面外科就诊的224例头颈部恶性肿瘤患者,均接受根治手术。根据LDVT发生情况分为发生组(n=24)和未发生组(n=200)。统计两组患者临床信息,通过单因素及多元Logistic回归筛选入组患者围手术期LDVT的独立危险因素,并构建风险预测列线图模型,采用受试者工作特征曲线(receiver operating characteristic,ROC)验证列线图模型的预测效能。结果两组在是否患有高血压、术前是否抗凝治疗、D-二聚体(D-dimer)水平、术后卧床时间和血小板与淋巴细胞比值(platelet/lymphocyte ratio,PLR)有明显差异(P<0.05)。将多元Logistic回归获取的危险因素构建风险预测模型,R语言软件计算模型的曲线下面积(area under curve,AUC)为0.814(95%可信区间0.778~0.849),提示其具有良好的判别和校准效果。决策曲线分析证实,在10%~75%的阈值概率区间内,预测模型的净收益较高。结论基于高血压病史、术前预防性抗凝治疗、D-二聚体水平≥0.5 mg/L、术后卧床≥3 d、血小板与淋巴细胞比值(PLR)≥176来构建的风险预测模型,对头颈部恶性肿瘤并接受根治手术的患者围手术期发生LDVT有良好的风险预测性,可用于提供个体化LDVT风险评估,指导治疗决策,减少LDVT并发症的发生。

成人股骨颈骨折术后股骨头坏死的多因素分析

目的 探讨成人患者股骨颈骨折术后发生股骨头坏死(ONFH)的相关危险因素。方法 回顾性分析2018年1月至2021年7月我院收治的150例股骨颈骨折患者的临床资料。根据患者是否出现ONFH分为坏死组和存活组。收集并比较2组患者性别、年龄、BMI、受伤后至手术时间、Garden分型、内固定是否取出、内固定取出时间、股骨头后倾角度、术后骨折复位质量等临床资料,并行单因素及非条件二分类Logistic回归分析。结果 150例成人股骨颈骨折患者中,41例出现术后ONFH(27.33%)。单因素分析结果显示,BMI、受伤后至手术时间、Garden分型、股骨头后倾角度、内固定取出时间以及术后骨折复位质量与ONFH的发生密切相关(P<0.05)。多因素Logistic回归分析结果显示,BMI≥25 kg/m^(2),Garden分型为Ⅲ、Ⅳ型,术后骨折复位质量为B、C级,内固定取出时间<2年是成人股骨颈骨折术后ONFH发生的独立危险因素。结论 成人股骨颈骨折术后ONFH的发生率较高。BMI、Garden分型、术后骨折复位质量、内固定取出时间均会影响股骨颈骨折术后ONFH的发生。临床应针对相关影响因素加强干预,以降低ONFH的发生率,提高患者生活质量。

以肺外为首次转移部位的头颈部腺样囊性癌患者的临床、预后特征及危险因素分析

目的总结分析以肺外为首次远处转移部位的腺样囊性癌患者的临床特征及预后情况,并探寻相关危险因素。方法回顾性分析并收集2002年1月~2020年12月就诊于北京同仁医院,病理明确诊断为腺样囊性癌患者126例,其中首次远处转移部位为肺外的患者21例,首次远处转移部位为肺的患者105例。此外,自美国国立癌症研究所监测,流行病学和最终结果(SEER)数据库中提取2010~2019年病理明确诊断为腺样囊性癌患者的临床数据进行预后分析。结果单因素分析显示N分级、是否伴随神经症状、病理组织学分型、分级、Ki67、神经侵犯、p63情况与肺外转移相关(χ²=5.385、9.888、20.485、15.579、8.711、5.476、5.280,P均<0.05)。多因素条件Logistic回归分析结果显示N分级、病理分级、神经症状与肺外转移风险增高存在相关性。生存分析显示首发肺外转移组的累积生存率、累积无进展生存率更低(P均<0.05)。结论腺样囊性癌首次转移部位与N分级、病理分级、神经症状多因素综合作用有关,实性型及伴高级别转化患者更易发生肺外转移。首次转移部位为肺外的患者远期预后较差。

基于logistic回归分析影响头颈部肿瘤患者睡眠的危险因素以及心理弹性变化轨迹

目的:探讨基于logistic回归分析影响头颈部肿瘤患者睡眠的危险因素以及心理弹性变化轨迹与睡眠之间的相关性。方法:回顾性分析2021年1月—2022年12月河南省肿瘤医院收治的850例头颈部肿瘤患者的临床资料。采用肿瘤心理弹性量表(RS-SC)和匹兹堡睡眠质量指数(PSQI)对患者的睡眠情况和心理弹性变化轨迹进行分析,并将PSQI得分>7分的患者纳入睡眠质量异常组(n=160),PSQI得分≤7分者纳入睡眠质量正常组(n=690)。分析影响头颈部肿瘤患者睡眠质量的风险因素。结果:放疗后,患者的RS-SC评分短暂升高,随着治疗时间的延长,RS-SC评分呈现逐渐降低的趋势,其PSQI得分整体呈先下降后上升的趋势。本研究纳入患者的睡眠质量异常率为18.82%;文化程度较低、放疗晚期、生理因素、环境因素、心理因素及频繁疼痛等因素均是影响头颈部肿瘤者睡眠质量异常的危险因素(OR=1.143、1.125、1.339、1.670、1.250、10.740,P<0.05)。结论:在头颈部肿瘤患者接受治疗期间患者心理弹性和睡眠质量均会发生明显改变,其中文化程度、放疗阶段、疼痛、心理、环境及生理因素均与其睡眠质量相关,应针以上风险因素为患者制定更为科学的干预方案。

基于深度学习算法的自动勾画系统在头颈部危及器官勾画精度的研究

目的:评估和分析3种基于深度学习技术的自动勾画系统在勾画头颈部危及器官(OAR)中的准确度。方法:以放疗医师手工勾画的OAR为标准,通过体积相似性系数(DSC)、豪斯多夫距离(HD)、感兴趣区域质心偏差(COMD)、过分割率(FNR)、欠分割率(FPR)、Jaccard系数(JC)、灵敏度指数(SI)及包容性系数(II)等参数评估PV-iCurve、RT-Mind和AccuContour自动勾画系统在头颈部OAR勾画的精度。结果:脑的FNR、JC、SI值,脑干的FPR、II值,左眼的FPR、FNR、JC、SI值,下颌骨的FPR、FNR、SI、II值,左腮腺的FPR、FNR、SI、II值及脊髓的DSC、FPR、JC、II值均显示3种勾画系统间存在统计学差异(P<0.05),只有脑干的HD、FNR、SI值和脊髓的HD值显示3种自动勾画系统勾画结果无统计学差异(P>0.05)。结论:通过多个参数的比较,发现3种软件在不同OAR勾画中的勾画精度不同,难以进行整体横向比较,因此这些参数仅作为参考,不能用于评估勾画结果作为临床治疗的标准,虽然3种软件都有较好的勾画结果,但仍需医师仔细审核和做必要的修改。

头颈癌病人身体意象障碍风险预测模型的构建及验证

目的:分析头颈癌病人发生身体意象障碍的危险因素,构建列线图预测模型,并验证模型的预测效果。方法:采用便利抽样法,选取2022年6月—11月在太原市某三级甲等医院耳鼻咽喉头颈外科、口腔科、放疗科住院的头颈癌病人以及2022年7月-9月在太原市某肿瘤专科医院头颈外科住院的共332例头颈癌病人作为研究对象,按照7∶3的比例分为建模组(235例)和验证组(97例)。依据治疗后是否发生身体意象障碍将建模组病人分为身体意象障碍组和身体意象正常组,通过Logistic回归分析探索身体意象障碍的危险因素,建立预测模型,绘制列线图,通过受试者工作特征(ROC)曲线下面积、Hosmer-Lemeshow拟合优度检验分别评价模型的区分度、校准度。结果:332例头颈癌病人中,身体意象障碍发生率为37.05%;根据肿瘤分期、有无气管套管、功能状况得分、焦虑得分、领悟社会支持得分、消极应对得分6个影响因素构建预测模型。该模型在建模组和验证组中的ROC曲线下面积分别为0.897(P<0.001)和0.850(P<0.001),区分度良好;Hosmer-Lemeshowχ^(2)值分别为7.828(P=0.348)和5.409(P=0.610),校准度较好。结论:构建的预测模型效果良好,可为临床医护人员评估头颈癌病人发生身体意象障碍的风险、进行早期干预提供参考。

股骨颈骨折空心螺钉内固定术后并发股骨头坏死风险列线图模型的建立与验证

目的 研究股骨颈骨折(femoral neck fracture, FNF)空心螺钉内固定(cannulated screw fixation, CSF)术后并发股骨头坏死的危险因素,构建一个个体化的股骨头坏死风险预测列线图模型。方法 回顾性分析作者医院骨科2018-09/2021-01月252例符合条件的FNF患者作为研究对象,并收集临床资料。通过单因素分析,筛选与CSF术后股骨头坏死相关因素,再将它们纳入多因素Logistic回归模型中进行危险因素分析。采用R软件建立列线图模型,并通过Bootstrap法对其进行内部验证并且绘制校准曲线;利用受试者工作特征(receiver operating characteristic, ROC)曲线及曲线下面积(area under the curve, AUC)评估预测模型的区分度,并绘制决策曲线分析(decision curve analysis, DCA)曲线,以评估预测模型的应用价值。结果 252例FNF患者CSF术后,61例并发股骨头坏死,发生率为24.21%。Logistic回归分析结果显示,年龄≥60岁、糖尿病史、头下型骨折、Garden分型Ⅲ~Ⅳ型、Pauwels分型Ⅱ型、VN分型Ⅱ~Ⅲ型为FNF患者CSF术后股骨头坏死的危险因素(P<0.05)。校准曲线预测FNF患者CSF术后股骨头坏死发生风险与实际发生风险的一致性较好,列线图模型预测FNF患者CSF术后股骨头坏死风险内部验证的ROC曲线AUC为0.968(95%CI:0.949~0.987),DCA表明当阈值概率在0~1时模型能够带来临床收益。结论 本研究成功构建了一个高精度的列线图预测模型,该模型包括年龄、糖尿病史、骨折线部位、Garden分型、Pauwels分型、VN分型等预测变量,有助于提高对FNF患者CSF术后并发股骨头坏死高危患者的早期识别和管理效率。

口腔颌面-头颈恶性肿瘤术后急性肾损伤的发生率及危险因素分析

目的:明确口腔颌面-头颈恶性肿瘤术后急性肾损伤(acute kidney injury,AKI)的发生率并分析相关危险因素。方法:回顾性分析上海交通大学医学院附属第九人民医院2014年1月—2021年12月接受口腔颌面-头颈恶性肿瘤切除手术的患者,收集围术期资料。根据改善全球肾脏病预后组织(Kidney Disease:Improving Global Outcomes,KDIGO)标准,术后48 h内血清肌酐升高≥26.5μmol/L或7天内血清肌酐升高≥1.5倍基础值定义为AKI。根据术后是否发生AKI分为2组—AKI组和非AKI组。统计术后AKI发生率,比较2组术后机械通气使用率、死亡率、ICU停留时间、住院时间和住院费用。采用SPSS 26.0软件包中的Logistic回归分析确定术后AKI的独立危险因素。结果:最终纳入6830例患者(男4238例,女2592例;平均年龄59.3岁)。134例(2.0%)患者术后发生AKI,其中AKI 1、2、3级分别为129例、4例、1例。多因素Logistic回归分析显示,女性(OR=2.166,95%CI:1.440~3.259)、非甾体抗炎药(non-steroidal anti-inflammatory drugs,NSAIDs)(OR=2.305,95%CI:1.088~4.883)、贫血(OR=1.688,95%CI:1.132~2.517)、蛋白尿(OR=2.080,95%CI:1.242~3.482)、长时间麻醉(OR=0.346,95%CI:0.196~0.612)是术后发生AKI的独立影响因素。AKI组患者在住院期间有更高的死亡率(1.0%),但AKI并不影响临床结局。结论:口腔颌面-头颈恶性肿瘤术后AKI的发生率约为2.0%,女性、NSAIDs、贫血、蛋白尿、长时间麻醉与AKI的发生独立相关。

基于深度学习的自动分割方法对头颈部危及器官勾画的评价

目的:探讨与评价AccuContour软件基于深度学习的自动分割方法对头颈部危及器官(OAR)自动勾画结果的准确性。方法:选取20例鼻咽癌患者的CT图像资料,应用基于深度学习的机器算法模型AccuContour软件,对图像的19个OARs(脑干、脊髓、双侧眼球、双侧晶体、双侧视神经、垂体、双侧腮腺、口腔、下颌骨、双侧颌下腺、双侧颞颌关节、甲状腺和喉)进行自动勾画,勾画结果定义为深度学习勾画的体积(V_(DL)),对其进行手动修改后,将修改结果定义为参考勾画(V_(ref)),采用相似性系数(DSC)、Jaccard系数(JAC)、平均Hausdorff距离(AHD)、体积差异比(ΔV)、最大剂量(D_(max))或平均剂量(Dmean)评价勾画结果。结果:除垂体、双侧颞颌关节及双侧视神经外,其他器官的DSC平均值均>0.7,JAC平均值均>0.6;各器官AHD均值最大为1.91 mm,最小为0.23 mm。双侧眼球、双侧晶体、下颌骨、双侧颌下腺、喉和甲状腺在剂量评估中表现优异,以ΔD_(max)评估的器官中,绝对平均值最大为5.81 Gy,以Dmean评估的器官中,绝对平均值最大为3.77 Gy。结论:AccuContour软件基于深度学习的自动分割方法在头颈部OAR自动勾画中能够获得较好的结果,可以用来提高临床工作效率。

股骨颈骨折患者空心螺钉内固定术后股骨头坏死的危险因素及其预测模型构建

目的分析股骨颈骨折(femoral neck fracture,FNF)患者空心螺钉内固定(cannulated screw fixation,CSF)术后股骨头坏死的危险因素,并构建FNF患者CSF术后股骨头坏死的联合预测模型。方法回顾性分析2019年1月—2021年6月保定市第二中心医院骨科收治的205例接受CSF术治疗的FNF患者临床资料。其中男性86例,女性119例;年龄28~70岁,平均52.7岁;致伤原因:道路交通伤、高处坠落等高能量损伤103例,跌倒、碰撞等低能量损伤102例;Garden分型:Ⅰ型65例,Ⅱ型39例,Ⅲ型58例,Ⅳ型43例。根据CSF术后1年是否发生股骨头坏死分为股骨头坏死组52例和无股骨头坏死组153例。采用单因素和多因素Logistic回归分析FNF患者CSF术后股骨头坏死的危险因素,构建FNF患者CSF术后股骨头坏死的联合预测模型,采用受试者工作特征(receiver operating characteristic,ROC)曲线评价预测模型的效能。结果患者均随访1年,205例FNF患者共发生52例股骨头坏死,股骨头坏死发生率为25.37%(95%CI 23.98~26.76)。单因素分析显示,两组年龄、吸烟、糖尿病、高血压、Garden分型、骨折至手术时间、复位质量和术中出血量比较差异有统计学意义(P<0.05)。多因素Logistic回归分析显示,年龄≥60岁、吸烟、糖尿病、高血压、Garden分型Ⅲ~Ⅳ型、骨折至手术时间≥48h、复位质量不满意为FNF患者CSF术后股骨头坏死的危险因素(P<0.05)。联合预测模型及ROC分析显示:其对FNF患者CSF术后股骨头坏死有较高的评估效能,曲线下面积为0.855(95%CI 0.780~0.925)。结论FNF患者CSF术后股骨头坏死的危险因素包括年龄≥60岁、吸烟、糖尿病、高血压、Garden分型Ⅲ~Ⅳ型、骨折至手术时间≥48h、复位质量不满意,通过危险因素构建的联合预测模型预测患者CSF术后股骨头坏死的价值良好。

头颈癌游离皮瓣患者术后谵妄危险因素及NLR的临床价值分析

目的:研究头颈癌患者游离皮瓣重建术后谵妄(POD)发生率和危险因素,构建风险预测模型,探讨中性粒细胞/淋巴细胞比值(NLR)是否具有成为头颈癌患者游离皮瓣重建术后谵妄生物标志物的潜力。方法:收集徐州市中心医院2016年1月—2022年3月行游离皮瓣重建的头颈恶性肿瘤住院患者128例,术后至出院每天根据《谵妄评估量表》评估患者是否谵妄并分为谵妄组和非谵妄组。采用SPSS 26.0软件包分析患者POD相关危险因素并用Logistic建立预测模型,以ROC曲线下面积检验模型预测效果及最佳截断值下的敏感度和特异度。结果:POD发生率为13.3%(17/128)。单因素分析显示,年龄≥65岁、术前NLR、术前总蛋白、手术时间、术中输血、术后ICU观察时间、术后睡眠紊乱、术后VAS评分具有统计学意义(P<0.05)。多因素Logistic回归分析显示,年龄≥65岁(OR=5.253;95%CI:1.146~24.074)、术前高NLR(OR=1.891;95%CI:1.050~3.405)、术中输血(OR=6.108;95%CI:1.109~33.644)、术后睡眠紊乱(OR=9.292;95%CI:1.441~59.914)、术后疼痛(OR=1.807;95%CI:1.018~3.206)是POD的独立危险因素。NLR-预测模型ROC曲线AUC=0.913,约登指数为0.522,NLR最佳截断值为2.540,灵敏度为0.765,特异度为0.757。结论:NLR-预测模型对头颈癌游离皮瓣重建患者POD预测价值良好,NLR具有生物标志物的潜力,可用于临床指导早期干预与治疗。

头颈癌切除游离皮瓣修复患者术后谵妄预测模型的建立与验证

目的建立并验证头颈部癌(head and neck cancer,HNC)患者游离皮瓣修复术后谵妄(postoperative delirium,POD)的风险预测模型,为临床提供参考。方法本研究已通过单位伦理委员会审查批准,并获得患者知情同意。对2016年1月1日至2022年1月1日在徐州市中心医院头颈肿瘤外科接受头颈癌切除后游离皮瓣重建术患者的相关危险因素和生命体征进行回顾性收集和评估,共纳入241例,其中171例进行预测模型的建立,另外收集70例进行内部验证,采用单因素和多因素Logistic分析,R Studio软件包进行建模统计分析。结果本研究模型最终包含4个风险因素:年龄、输血、术后睡眠障碍、术后疼痛VAS得分;该模型训练集的受试者工作特征曲线下面积(AUC)为0.869(95%CI:0.789~0.948),Youden指数为0.692,预测值为0.215,敏感性为85.3%,特异性为83.9%,Hosmer-Lemeshow拟合优度为10.336(P=0.242),该模型拟合较好;根据模型对验证集进行验证,C指数为0.827(95%CI:0.681~0.973),模型预测效果较好。结论该模型可应用于头颈部恶性肿瘤切除后游离皮瓣术后谵妄的预测,对入院时HNC-POD的风险具有很高的预测价值,使用该模型可能有助于更好地实施预防性治疗和护理措施。