首页/行业解读/我们用两种人工智能,来审阅一篇来自台湾的骨科论文,发现.../
我们用两种人工智能,来审阅一篇来自台湾的骨科论文,发现...
2024-01-17 10:58:14250浏览
我们用两种人工智能,来审阅一篇来自台湾的骨科论文,发现...

本公众号近期发布了一些AI阅读论文的文章。其实AI的一大好处是,不但可以辅助阅读论文,还能辅助审查和修改论文!比如,“骨科医生爱科技”知识星球赠送的“骨科资源聚合”就内置了一些AI资源,另外还内置了SCI论文写作辅助的一些链接。其中最重要的是“翻译”和“审稿”两个功能。

今天,我们用一篇台湾省骨科医生的论文来测试一下,AI作为“审稿人”,水平怎么样。

插一句,最近台湾省搞“选举”,其中柯文哲和赖清德都是医生出身,柯文哲是搞重症的,业务水平很高,是台大医学院教授级别;赖清德是内科医生出身,不过很早就从政了。

我们选择了一篇2023年6月发表的一篇骨科论文,主要对比了股骨远端粉碎性骨折的双钢板治疗相比于单钢板可以更好地预防术后内翻畸形。

论文摘要如下:Abstract 摘要Introduction 引言Varus collapse followed by osteosynthesis for distal femoral fractures with conventional implants has been well documented but is seldom mentioned in fractures managed with locking plates. The purpose of this study was to assess the incidence of varus collapse after treating complex supra-intercondylar fractures of the distal femur (AO type C3) using a Single Plate (SP) or Double Plate (DP) fixation technique.使用常规植入物治疗股骨远端骨折发生内翻塌陷的情况已有充分文献记载,但在使用锁定钢板治疗的骨折中很少提及。本研究的目的是评估使用单板 (SP) 或双板 (DP) 固定技术治疗股骨远端复杂髁间上骨折(AO C3 型)后内翻塌陷的发生率。

Materials and methods 材料与方法We retrospectively reviewed 357 patients with distal femoral fractures who were treated at our hospital between 2006 and 2017. After excluding cases of infection, malignancy, periprosthetic fracture, revision surgery, pediatric fracture, and extra-articular fracture, 54 patients were included in the study. All demographic data and radiological and clinical outcomes were reviewed and analyzed.我们回顾性研究了 2006 年至 2017 年间在我们医院接受治疗的 357 例股骨远端骨折患者。在排除感染、恶性肿瘤、假体周围骨折、翻修手术、小儿骨折和关节外骨折病例后,该研究纳入了 54 例患者。对所有人口统计学数据以及放射学和临床结局进行了审查和分析。

Results 结果There were 54 patients enrolled into this study with age from 15 to 85 years old (mean 41.6, SD = 19.9), and 32 of them were open fractures (59%). The patients were further divided into either an SP (n = 15) or a DP group (n = 39). Demographics, including age, sex, injury severity score, and open fracture type, were all compatible between the two groups. The overall nonunion rate was 25.9% (n = 14; 6 from the SP and 8 from the DP group; p = 0.175). The varus collapse rate was 9.3% (n = 5; 4 from the SP and 1 from the DP group (p = 0.018).本研究共有 54 名患者,年龄在 15 至 85 岁之间(平均 41.6,SD = 19.9),其中 32 例为开放性骨折 (59%)。患者进一步分为 SP 组 (n = 15) 或 DP 组 (n = 39)。两组之间的人口统计学特征,包括年龄、性别、损伤严重程度评分和开放性骨折类型,均相符。总体不愈合率为25.9%(n=14;SP组为6,DP组为8;p = 0.175)。内翻塌陷率为9.3%(n=5;SP组为4,DP组为1(p = 0.018)。Conclusions 结论The varus collapse rate after osteosynthesis with a single lateral locking plate could be as high as 26.7% in AO type C3 fractures of the distal femur, which would be decreased to 2.6% by adding a medial buttress plate. Surgeons should consider DP fixation to avoid varus collapse in severely comminuted complete intra-articular fractures of the distal femur.在股骨远端AO型C3骨折中,单侧锁骨板接骨术后内翻塌陷率可高达26.7%,增加内侧支撑板可降低至2.6%。外科医生应考虑 DP 固定,以避免股骨远端严重粉碎性完全关节内骨折发生内翻塌陷。

从摘要来看,这个论文结构不复杂。我们用ChatGPT和Claude分别审稿,得到的审稿意见如下

01ChatGPT的审稿意见

见下图。我们可以看到审稿意见还是很牛的,起码超过我们的很多身边的同事。其实,有研究表明,ChatGPT充当审稿人,和业内专家有一拼。审稿意见不但提出了文章的新颖性和创新性,还给出了接受稿件的可能原因、拒绝稿件的可能原因,以及修改的意见和建议。

02claude的审稿意见

我们可以看到,claude的审稿意见,比起ChatGPT-4来差一点。但claude也有它的优势,比如免费、可以处理多文档和长文本等。

友情链接: