As science has advanced and new diseases have been identified, proposed classifications have become ever more complex, and informal division of the medical fraternity has occurred into those who are termed “lumpers” and those who are “splitters.” From the splitter's viewpoint, the lumper is someone who dislikes change and sees no reason to employ subdivisions where there is no practical value from his or her own point of view. From the lumper's viewpoint, the splitter is someone obsessed with subdivision, working to promote his/her own academic existence by redrawing boundaries, gathering air miles, and publishing at wearisome length and in multiple places thereafter. Taken to its extreme, the lumper would view all patients as dead, sick, or well, whereas a splitter would believe every individual has a different disease.

In general, pathologists “split” as they delve into ever greater detail, whereas physicians tend to “lump” as treatment options often lag behind proposed new divisions. Congenital cysts in the lung are a good example of how opinion on current proposed classifications remains divided (1–4). Furthermore, different specialties frequently use different systems for the same diseases. Small airway disease is an example in which clinicians classify primarily according to cause, radiologists according to indirect versus direct signs on computed tomography (CT) (5), and pathologists on microscopic morphology (acute/chronic, proliferative/obliterative), as very rarely will there be features diagnostic of a specific cause (6). Yet they are all talking about the same diseases.

However, in 2002, a classification system was proposed for adult diffuse parenchymal lung disease that was not just “horizontal” (i.e., purely pathologic, radiological, or clinical) but incorporated a “vertical” component whereby the proposed classification was a dynamic interactive process involving all three disciplines (7). This internationally conceived classification system has led to much more ordered management of idiopathic interstitial pneumonias, with subsequent papers documenting its reproducibility (8, 9) and more controlled trials into therapy for idiopathic pulmonary fibrosis. Hence, the lumper's term of “cryptogenic fibrosing alveolitis” as a specific clinical disease is now effectively redundant (10).

The study and classification of pediatric diffuse lung disease (DLD) has lagged behind that of adults for a number of reasons: childhood interstitial lung diseases are at least an order of magnitude rarer (11), there is a wider range of diagnoses, pediatricians may be more reluctant to resort to invasive procedures, and diseases occur in the context of the developing lung. In this issue of the Journal (pp. 1120–1128), Deutsch and colleagues propose a new classification for DLD in infants, again incorporating a multiinstitutional and multidisciplinary approach (12). The work is impressive, not least in that it brings together data from a cohort of infants all of whom have rare and difficult to diagnose disorders. Even the most committed lumper would agree that the holistic view of calling all such cases “fibrosing alveolitis in children” is no longer tenable and, although the pathologic terminology is sometimes the same as in adults, the clinical correlates are different. An example is desquamative interstitial pneumonia, which is usually associated with exposure to cigarette smoke in adults, but is mainly due to gene mutations in children (12, 13). The classification also incorporates disease processes unique to this age group (diffuse neuroendocrine cell hyperplasia of infancy [NEHI] and pulmonary interstitial glycogenolysis [PIG]), and the integration of molecular studies into diagnostic practice for DLDs is a notable step forward.

There are other differences between the two age groups in that high-resolution CT in the pediatric age group is far less sensitive or specific than in adults, so it will have less impact on the decision to biopsy; conversely molecular tests may preclude the need to biopsy in other cases. This again serves to emphasize the increasing importance of a multidisciplinary approach.

A new classification inevitably brings new questions. From the lumper's perspective, is PIG a specific disease or simply a morphologic marker of lung immaturity within the interstitium, which may predominate in some samples? Is NEHI a disease or a secondary event? Can growth and development really be separated into discrete categories, when they are evidently parallel processes? From the splitter's perspective, how will this classification deal with DLDs that generally do not come to lung biopsy—for example, idiopathic pulmonary hemosiderosis? Finally, independent of lumpers and splitters, one could argue that this proposed classification of DLDs is too broad: for example, should the classification encompass chronic lung disease of prematurity, which is usually a management rather than a diagnostic conundrum?

Such considerations aside, this proposed classification system provides a system through which cases can be diagnosed and collected in sufficient number for further analysis, providing a template for potential better clinical management. In adults, concentrating reporting into specialist multidisciplinary teams improves diagnostic accuracy (14), and so management of infants with these diseases will hopefully follow suit. Furthermore, the pathologic documentation by Deutsch and colleagues is a model of its kind, and will hopefully lead to clinicians following pathologists' example in developing protocols (11, 15), prospectively collecting both clinical information and data on noninvasive investigations in a systematic way, and using common CT protocols with reporting by specialist radiologists. There is exactly no chance of making progress with therapeutics in these rare conditions, unless networks are established that categorize patients in a uniform manner, and that conduct randomized trials of therapies. This will be particularly important as more cytokine-based therapies come on line; if they are applied indiscriminately to “lumped” groups of patients, then potentially beneficial therapeutic effects will be missed, children will be put at risk unnecessarily, and money will be wasted.

This classification should be prospectively tested for diagnostic reproducibility and against other disease cohorts outside the United States, and then refined accordingly. Whether those children aged from 2 to 18 years of age require a separate classification remains to be seen, but arguments over which disease groups should be considered ineliminable, and which should not, will no doubt continue (16).

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES