Hi Linh,
I think as a general rule we can not always be completely specific and this is a common theme with ultrasound for example solid breast lesions and ovarian lesions.
But as in these examples we can give some guidance,
With cellulitis I expect to see subcutaneous oedema and hyper-vascularity with no discreet collection. If it is in the lower limb you may note some reactive groin nodes and a hyper dynamic superficial venous return to build your diagnosis. The clinical appearance of shiny red skin also helps (never look at the ultrasound in isolation, always take the clinical presentation into account).
With a phlegm you will see an inflamed area and centrally within it there will be an area of reduced vascularity compared to the hyper-vascularity of the infective surrounding tissue. This indicates some breakdown heading towards abscess formation. If you compress this tissue you will not see much tissue movement and I would conclude it is a phlegm.
In an abscess the general tissue is inflamed and hypervascular and in the middle of this tissue where the abscess has formed you will have an avascular area. If you compress it (squash it) wit the transducer you may see the puss move within the abscess. I call this my “+ve for sloshy sign” I use the same technique when I am looking at hydroxyapatite deposits in the MSK system to see if they are aspiratable.
A seroma is a post operative collection which is filled with serous material. The fluid within is usually anechoic (if it is a haematoma or abscess the fluid will have internal echoes) The wall is usually thin and the internal contents are avascular. If you drain these they may re-accumulate.Another interesting type of serum is the Morel Lavallee lesion in the MSK system where the superficial and deep fascias of the sub-cutaneous fat slide agains each other during trauma and a seroma results full of blood and lymphatic fluid. Once again these often re-accumulate following aspiration.
If it is a haematoma it will have some internal echoes and may have some clot forming within or separation of the blood components (plasma Vs red blood cells).
Granulation tissue is scar tissue, so it is solid, hypo echoic, firm on palpation, avascular or hypo vascular and in the expected location following surgery or trauma.
So in summary I think these entities have a variety of appearances that are characteristic of the correct diagnosis but perhaps not specific, with some overlap.
Steve