SEE EDITS BELOW. Keeping original post to further demonstrate bad journalism.

The weirdness of the question made me look up the news article and journal article itself, and after reading it I'm not sure the journal article says anything of the sort.

In our sample of 20 individuals from Londinum, we suggest that four people had migrated from outside of the settlement and that twelve people were either born in and/or grew up in the immediate Roman London area. The origins of the remaining four individuals are less clear. Our results lend further weight to the results of other isotopic studies addressing origin, cultural identity, and funerary practice in Roman Britain, where there is not always a direct correlation between these variables.

So 12 are from Londinum, 4 are not, and 4 are unclear.

Of the 4 known to be from outside Londinium:

MSL87-390 (36-45 female)

This is a unique burial in Londinium and the style of dress signifies a non-local identity that is in keeping with her Sr isotope profile. Her Pb isotope composition, however, is within the range of a Roman British origin and may either be suggesting that this individual originated from somewhere within Roman Britain other than Londinium (based on her Sr isotope composition), but has a strong cultural affiliation with Germany; or that this individual originated from a region of the Roman Empire (possible Germany) with an anthropogenic Pb composition similar to Roman Britain

BGB98-400 (8 year old)

A concentration of 4.4 ppm suggests the Pb exposure was geogenic exposure and the subadult may have originated in an area near the Rhine Valley of Germany where both the Sr and Pb isotope compositions could be accommodated.

Also her grave goods and burial method suggested she was local.

GDV96-325 (18-25 female)

She has a high tooth enamel Pb concentration (10 ppm) and plots within the field of circum-Mediterranean anthropogenic Pb isotope composition.

SRP98-34245 (old adult male)

This individual (SRP98-34245) has Pb isotope ratios consistent with the area around Rome (Italy). He has a very high Sr concentration (beyond the 2SD range for British tooth enamel) of 268 ppm and a Sr isotope composition that would be consistent with limestone regions around the Mediterranean.

Of the 4 inconclusive results:

The adult male HOO88-652 has Pb isotope results that are suggestive of a non-local origin, but not conclusively... GDV96-150 is the high status burial of a seven-year-old child, who was buried with a glass flask, jet pin, hobnails, and accompanying chicken burial. This individual's Sr composition is consistent with a childhood spent in Londinium, but had the highest Pb concentration at 14.6 ppm of the entire sample set and a Pb isotope ratio that falls on the edge of the English ore Pb field, possibly suggesting a non-local origin.... LOW88-803.6 is a cranium recovered from a pit outside the city walls and has a Pb isotope composition that plots in an area where the Roman coin data field and the Pb isotope composition of the Mendip Ore field [Somerset] overlap, but are outside the field of anthropogenic English Pb. This male displays the lowest Pb ppm concentration of all twenty samples at 0.24 ppm which strongly suggests no, or minimal, exposure to anthropogenic sources of Pb during childhood. He also has a slightly elevated Sr isotope composition (0.71034), which would support a non-Londinium origin, although, cannot independently rule out a childhood spent elsewhere in Britannia... LOW88-695.5 was recovered from as the same site as LOW88-803.6 and has a Pb isotope composition that is on the edge of the anthropogenic English Pb ore field. Additionally, this individual has one of the lowest Pb concentrations at 1 ppm. Although this individual could be non-local, as with GDV96-150 above, he falls too close to the culturally defined group to definitively exclude him being from Roman Britain. Moreover, the Sr isotope composition for this individual, however, is compatible with a Londinium origin. However, as mentioned previously, the Sr isotope ratios for the London area are shared with many other parts of Europe.

So of these four, it is unknown if they are locals of Londinium, Britain, or other areas of Europe.

So not one of the 20 are from areas outside of the Roman Empire.

Source: Shaw, Heidi, et. al. “Identifying migrants in Roman London using lead and strontium stable isotopes.” Journal of Archaeological Science, Volume 66. February 2016.

That’s the article linked to from this news article. So either the journalist cited and linked the wrong article, or she didn’t read it.

EDIT 1: She linked the wrong article. I found the right one after going to dailymail and then the times (neither of which named or linked the article smh).

The presence of individuals with African ancestry is not unexpected, as studies from other urban centres in Roman Britain have found people with this ancestry and/or mobility isotope result and our results combined with the funerary evidence for these individuals, supports the findings of Eckardt and colleagues who show that childhood origin and ancestry are not always reflected in a person's burial identity. The people with Asian ancestry are the first to be reported from Roman Britain however, this is cautiously asserted because the method we applied is based on more recent populations, and we recognise that morphology is subject to temporal and spatial variation. To the best of our knowledge, the only other individual from the Roman Empire identified as having Asian ancestry, is an adult male buried at the Imperial estate of Vagnari who was not local to the area and whose mtDNA revealed east Asian affiliations. It may well be that these individuals were themselves or were descended from enslaved people originating from Asia, as there were slave-trade connections between India and China, and India and Rome. These results will be further explored using aDNA work.

From Redfern, Rebecca, et. al. “Going south of the river: A multidisciplinary analysis of ancestry, mobility and diet in a population from Roman Southwark, London.” Journal of Archaeological Science, Volume 74. October 2016.

So the methodology is skeletal morphology. And it's based on modern people using their ancestry circa 19th century (to identify 2~4th century people). The author itself cites this as a weakness.

The types are taken from the US National Institute of Justice. A further weakness I can see is that the morphology groupings are only based on three type: African--based on East and West African, Asian--based on Chinese and Japanese, and European--based on German, Holland, and American white. There's no talk of morphology of Eastern Europeans, Persians, steppe nomads, etc. I'd like to see an analysis of difference between Chinese, Persian/Middle-Eastern, Mongol/Tartar, and European skeletons before I'm comfortable of calling these individuals Chinese.

So please wait for DNA and further lab work. The author also notes that ancestry could be an explanation.

EDIT 2: Not much data was laid out for the oxygen isotope analysis for some reason. That's the method the paper used to track mobility over an individual's life time. We're not told of the isotype analysis results for the individuals with Asian skeletal morphology, only the individuals with African skeletal morphology (and between 5 and 8 of 19 others that we're not told whom) spent their childhood in warmer climates. That kind of implies in my mind that the individuals with Asian skeletal morphology don't especially stand out in terms of oxygen isotope analysis, or they would've mentioned it. But I wouldn't know because the paper did not go into detail and list out everyone's oxygen isotopes. So if (and so far that is a huge if) the identification of Asians based on skeletal morphology is correct, it's probably Asian descent, not someone who actually traveled from East Asia. That's my guess based on the data presented.