Purpose

Many studies have shown that patients with non-central nervous system (CNS) cancer can have brain abnormalities, such as reduced gray matter volume and cerebral microbleeds. These abnormalities can sometimes be present even before start of treatment, suggesting a potential detrimental effect of non-CNS cancer itself on the brain. In these previous studies, psychological factors associated with a cancer diagnosis and selection bias may have influenced results. To overcome these limitations, we investigated brain structure with magnetic resonance imaging (MRI) prior to cancer diagnosis.

Patients and methods

Between 2005 and 2014, 4,622 participants from the prospective population-based Rotterdam Study who were free of cancer, dementia, and stroke, underwent brain MRI and were subsequently followed for incident cancer until January 1st, 2015. We investigated the association between brain MRI measurements, including cerebral small vessel disease, volumes of global brain tissue, lobes, and subcortical structures, and global white matter microstructure, and the risk of non-CNS cancer using Cox proportional hazards models. Age was used as time scale. Models were corrected for e.g. sex, intracranial volume, educational level, body mass index, hypertension, diabetes mellitus, smoking status, alcohol use, and depression sum-score.

Results

During a median (interquartile range) follow-up of 7.0 years (4.9–8.1), 353 participants were diagnosed with non-CNS cancer. Results indicated that persons who develop cancer do not have more brain abnormalities before clinical manifestation of the disease than persons who remain free of cancer. The largest effect estimates were found for the relation between presence of lacunar infarcts and the risk of cancer (hazard ratio [HR] 95% confidence interval [CI] = 1.39 [0.97–1.98]) and for total brain volume (HR [95%CI] per standard deviation increase in total brain volume = 0.76 [0.55–1.04]).

Conclusion

We did not observe associations between small vessel disease, brain tissue volumes, and global white matter microstructure, and subsequent cancer risk in an unselected population. These findings deviate from previous studies indicating brain abnormalities among patients shortly after cancer diagnosis.

中文翻译：

---

非中枢神经系统癌症诊断前的大脑结构：一项基于人群的队列研究

目的

许多研究表明，患有非中枢神经系统（CNS）癌症的患者可能患有脑部异常，例如灰质减少和脑微出血。这些异常有时甚至在治疗开始之前就可能存在，表明非CNS癌症本身对大脑可能具有有害作用。在这些先前的研究中，与癌症诊断和选择偏见相关的心理因素可能会影响结果。为了克服这些限制，我们在癌症诊断之前通过磁共振成像（MRI）研究了大脑结构。

患者和方法

在2005年至2014年之间，来自于鹿特丹的前瞻性人群研究的4,622名无癌，痴呆和中风的参与者进行了脑MRI检查，随后随访其发生的癌症直到2015年1月1日。我们调查了脑MRI测量值之间的关联使用Cox比例风险模型，包括脑小血管疾病，整体脑组织，叶和皮层下结构的体积以及整体白质微结构，以及非中枢神经系统癌症的风险。年龄用作时间尺度。对模型进行了校正，例如性别，颅内体积，教育水平，体重指数，高血压，糖尿病，吸烟状况，饮酒和抑郁总和。

结果

在7.0年（4.9-8.1）的中位（四分位间距）随访中，有353名参与者被诊断出患有非中枢神经系统癌症。结果表明，罹患癌症的人比没有癌症的人没有更多的脑部异常。发现最大的影响估计是腔隙性梗塞的存在与癌症风险之间的关系（危险比[HR] 95％置信区间[CI] = 1.39 [0.97-1.98]）和总脑容量（HR [95％]） CI]每标准偏差的增加使总脑容量= 0.76 [0.55-1.04]）。

结论

我们未观察到小血管疾病，脑组织量和整体白质微观结构与未选定人群的后续癌症风险之间的关联。这些发现与先前的研究结果有所不同，先前的研究表明癌症诊断后不久患者中脑部异常。