甲狀腺癌自然疾病史和預後之類病例世代研究

A Non-standard Case-cohort Study for the Disease Natural History and Prognosis of Thyroid Cancer

蘇登煌(Deng-Huang Su)

指導教授 ： 陳秀熙

共同指導教授 ： 張天鈞(Tien-Chun Chang)

國立臺灣大學/公共衛生學院/預防醫學研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.03270

摘要

研究背景與目的
甲狀腺癌一直是內分泌科中最重要的癌症，其發生率在近年來也是一直在增加，但國內對這個主題仍缺乏重視。雖然大部份的甲狀腺癌預後都很好，但仍有一部份的患者會面臨死亡，什麼因子與其存活有關呢?過去一直都是用存活分析來處理這個問題，有無其他的方法可以解決這個問題呢?而甲狀腺超音波用在甲狀腺癌的偵測上，一直沒有確鑿的證據來證明其對甲狀腺癌患者的存活有助益，是否可以其他方法嘗試來解決這個問題?本論文的目的在於建構一個甲狀腺癌的世代，透過這個世代的建立來建構甲狀腺癌的自然史及多階段的治療史，以了解不同的因子作用，並以不同模式之間的比較來估計使用甲狀腺超音波在甲狀腺癌高危險群的篩檢其可能之功效。

研究方法
本研究為以醫院為背景的回溯性研究，採類個案世代研究(non-standard case-cohort design)的方法進行，將西元1985年後在台大醫院進行甲狀腺超音波檢查的民眾及甲狀腺癌患者(依死亡及存活分類)，依不同比例抽樣來選取個案，使用馬可夫多階段模式構成甲狀腺癌多階段模式。先建構多階段自然史模式，並運用單一變數分析及多變數分析了解各因子是否為有影響之因子及其參數。再建構多階段治療史模式，並運用單一變數分析及多變數分析了解各因子的作用，並估算使用甲狀腺超音波的介入對於模式的影響。

主要研究結果
在甲狀腺癌自然史的分析，其發生率主要還是受到性別和年齡的影響，並且存在著性別和年齡的交互作用；當甲狀腺癌在無法以手術治愈的過程中，其相關因子為診斷時年齡大於60歲、腫瘤大於1公分以上、有侵犯甲狀腺以外的組織及甲狀腺癌組織分類(濾泡癌和髓質癌)。另在治療後轉成無法以手術治愈的分化甲狀腺癌的過程中，診斷時年齡大於40歲以上、腫瘤大於1公分以上及侵犯甲狀腺以外組織為此階段之預後因子。當其轉成無法以手術治愈的分化甲狀腺癌後而死於此癌症的過程中，診斷時年齡大於40歲以上、濾泡癌和髓質癌為其預後因子。當轉變成未分化癌或分化不良的甲狀腺癌之過程中，有男性、診斷時年齡大於60歲和侵犯甲狀腺以外組織等預後因子。另在未分化癌或分化不良的甲狀腺癌死於甲狀腺癌的過程中，診斷時其年齡介於40-60歲為預後因子。
甲狀腺超音波檢查對於預防甲狀腺癌是否進展至較差的階段有13%(95%CI: -25%~41%)的功效，其主要的功效是在預防分化良好的甲狀腺癌變成未分化癌或分化不良的甲狀腺癌。

結論
本研究應用類個案世代研究於甲狀腺癌自然史及預後因子之探討，為有效又具經濟之研究設計，除可節省大量的時間和人力，而之所建構之甲狀腺癌多階段模式也確實可找出預後因子及評估甲狀腺超音波介入的功效，而將甲狀腺癌用多階段模式來了解其自然病程，目前是沒有文獻採類似的方法，本研究是最先採用此模式之研究。

關鍵字

甲狀腺癌 ； 預後 ； 甲狀腺超音波 ； 馬可夫多階段模式 ； 個案世代研究 ； 甲狀腺癌自然疾病史

並列摘要

Background and purpose
The incidence rate of thyroid cancer has been increasing in recent years. While the prognosis of thyroid cancer is good, but there are still a fraction of the patients that eventually have poor prognosis. Thyroid ultrasound has been adopted for early detection of thyroid cancer, but the evidence is elusive. The aims of this study were to build a cohort of thyroid cancer and construct the multi-state natural disease history and multi-state treatment history of thyroid cancer to assess how natural progression of thyroid cancer evolves and how putative personal attributes and clinical attributes affect its natural course. Efficacy of thyroid ultrasound was further estimated by using two kinds of stochastic process.

Methods
This study was based on a hospital-based, non-standard case-cohort design. The cohort was composed of the patients with thyroid ultrasound in National Taiwan University Hospital after1985 and patients with thyroid cancers. The cases were randomly selected by different sampling ratios according to the status of thyroid cancer and survival. Multi-state Markov models of natural diseaase history and treatment history of thyroid cancer were constructed. Bayesian inversion tailor for non-standard case-cohort design was applied to identify significant factors responsible for each transition. The estimate of efficacy of thyroid ultrasound was done between the comparisons of different models. .

Main findings
In the natural disease history of thyroid cancer, the annual incidence rate from free of thyroid cance to ETC (echo-detectable thyroid cancer) was mainly affected by gender and age. Factors responsible for the transition between ETC and incurable stage included age of diagnosis, tumor size larger than 1 cm, the invasion of the extrathyroidal tissue, and the classification of thyroid cancer (medullary and follicular carcinoma). The transition between ETC and incurable differentiated thyroid cancer after treatment was affected by age of diagnosis, tumor size larger than 1 cm, and the invasion of the extrathyroidal tissue. Prognostic factors related to death from incurable differentiated thyroid cancer included age of diagnosis, medullary carcinoma and follicular carcinoma. Significant relevant factors pertaining to an anaplastic carcinoma or poorly differentiated thyroid cancer were male, age of diagnosis older more than 60 years old and the invasion of the extrathyroidal tissue. Significant factor for the transition anaplastic cancer or poorly differentiated thyroid cancer to death was was age of diagnosis between 40-60 years old. The efficacy of the ultrasound was 13% (95%CI:-25
-41%) in the arresting the procession from ETC to incurable thyroid cancer, and the efficacy was mainly in the prevention of well-differentiated thyroid cancer into anaplastic cancer or poorly differentiated thyroid cancer.

Conclusion
Non-standard case-cohort study together with a multi-state model is efficient in elucidating the natural history of thyroid cancer, significant factors related to natural course, and the efficacy.

並列關鍵字

Thyroid cancer ； Prognosis ； Thyroid ultrasound ； Markov multi-state model ； Case cohort study ； Thyroid cancer natural disease history

參考文獻

1. Cappelli C, Castellano M, Pirola I, Gandossi E, De ME, Cumetti D, Agosti B, Rosei EA. Thyroid nodule shape suggests malignancy. Eur J Endocrinol 2006;155:27-31.

3. Hoang JK, Lee WK, Lee M, Johnson D, Farrell S. US Features of thyroid malignancy: pearls and pitfalls. Radiographics 2007;27:847-60.

4. Carpi A, Nicolini A, Casara D, Rubello D, Rosa PM. Nonpalpable thyroid carcinoma: clinical controversies on preoperative selection. Am J Clin Oncol 2003;26:232-35.

5. Ron E, Lubin JH, Shore RE, Mabuchi K, Modan B, Pottern LM, Schneider AB, Tucker MA, Boice JD, Jr. Thyroid cancer after exposure to external radiation: a pooled analysis of seven studies. Radiat Res 1995;141:259-77.

14. Kim EK, Park CS, Chung WY, Oh KK, Kim DI, Lee JT, Yoo HS. New sonographic criteria for recommending fine-needle aspiration biopsy of nonpalpable solid nodules of the thyroid. AJR Am J Roentgenol 2002;178:687-91.